Netezza Data Loading Guide PDF

IBM Netezza 6.0.
IBM Netezza Data Loading Guide

Revised: June 5, 2012
20525 Rev. 3
Note: Before using this information and the product that it supports, read the information in “Notices and Trademarks” on
page D-1.
© Copyright IBM Corporation 2011, 2012.

US Government Users Restricted Rights – Use, duplication or disclosure restricted by GSA ADP Schedule Contract with IBM
Corp.
Table of Contents
Preface
1 Overview
Data Loading Components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-1
Data Loading Formats . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-2
New Decimal Delimiter Option . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-2
2 External Tables
About External Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-1
Privileges Required . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-2
Displaying External Table Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-2
Log Files. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-2
Usage. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-2
Parsing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-3
Backing Up and Restoring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-4
Command Syntax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-4
Transient External Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-4
Syntax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-4
Explicit Schema Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-5
Implicit Schema Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-5
Exporting Data Using Transient External Tables . . . . . . . . . . . . . . . . . . . . . . . . . 2-5
Remote Transient External Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-5
Supported Data Types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-6
Integer Data Types. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-7
Fixed-Point Data Types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-7
Floating-Point Data Types. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-8
Character Strings. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-10
Time Data Types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-11
Restrictions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-13
Best Practices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-13
Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-15
Transient External Table. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-15
Fixed-Length Format . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-16
Standard Unloading and Reloading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-16
iii
Back up and Restore a User Table. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-17
3 External Table Options

Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-1
Option Details . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-3
BoolStyle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-3
Compress . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-4
CRinString . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-4
CtrlChars . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-4
DataObject . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-4
DateDelim . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-5
DateStyle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-5
DecimalDelim . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-6
Delimiter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-6
Encoding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-7
EscapeChar. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-7
FillRecord. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-8
Format . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-8
IgnoreZero . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-8
IncludeZeroSeconds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-8
Layout . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-9
LogDir . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-9
MaxErrors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-9
MaxRows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-9
NullValue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-9
QuotedValue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-10
RecordDelim . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-11
RecordLength . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-11
RemoteSource. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-11
RequireQuotes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-11
SkipRows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-11
SocketBufSize. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-12
TimeDelim . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-12
TimeRoundNanos . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-12
TimeStyle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-12
TruncString. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-12
Y2Base. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-13
iv
Option Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-13
Counting Rows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-13
Handling Bad Rows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-14
Delineating Input Rows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-14
Matching Input Fields to Table Columns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-14
Using String and Non-string Fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-14
Handling the Absence of a Value. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-14
Enabling Load Continuation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-15
Handling Legal Characters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-15
Session Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-16
4 Using nzload
How the nzload Command Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-1
Protection and Privileges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-1
Concurrency and Transactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-2
Program Invocation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-2
Using the nzload Command . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-2
Syntax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-2
Inputs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-3
Additional Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-3
Outputs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-4
Using a Control File . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-5
Configuration File Example. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-6
5 Unloading Data
Unloading Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-1
Unloading Data to a Remote Client System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-2
6 Using Fixed-Length Format

Formatting Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-1
Fixed-Length Format . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-1
Data Attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-2
Format Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-2
New Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-2
Changed Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-3
Unsupported Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-3
Default Values. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-4
Layout Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-4
v
Building the Fixed-Length Format Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-6
End-of-Record. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-6
Record Length . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-6
Skipping Fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-6
Temporal Values . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-6
Numeric Values . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-7
Logical Values . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-8
Null Values . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-8
Appendix A: Examples and Grammar

The nzload Command . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-1
Specifying nzload Arguments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-1
Using Named Pipes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-2
Sample nzload Usage. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-2
Reference Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-4
Decimal Delimiter Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-4
SQL Grammar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-5
Fixed-Length Format Definition. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-7
Script Example For Loading Data Using Fixed-Length Format . . . . . . . . . . . . . . . . . . A-9
Appendix B: Troubleshooting
Tips for Successful Loading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B-1
Create Your Table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B-1
Determine Your Data Format. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B-1
Consider the Load Source. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B-2
Run the Job . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B-2
Troubleshoot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B-3
Handle Exceptions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B-3
Validate the Results. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B-3
Generate Statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B-3
Test Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B-3
nzload Error Handling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B-4
Reporting Errors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B-4
Understanding nzload Log Files . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B-4
Appendix C: Option Names

Specifying Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C-1
Appendix D: Notices and Trademarks
vi
Notices. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D-1
Trademarks. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D-3
Electronic Emission Notices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D-4
Regulatory and Compliance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D-7
Index
Rev. -vii
-viii Rev.
Preface
The IBM Netezza Data Loading Guide describes the functionality for data loading.
Audience for This Guide

The IBM Netezza Data Loading Guide is written for administrators using data loading
features.
Purpose of This Guide

This guide contains the following information:
Topics See the following
Introduction to Data Loading Concepts and Terms Chapter 1, “Overview”
How to use External Tables Chapter 2, “External Tables”
External Table options to use, and how the sys- Chapter 3, “External Table Options”
tem processes them
Details on the nzload command Chapter 4, “Using nzload”
Details on unloading data Chapter 5, “Unloading Data”
Details on the Fixed-Length format Chapter 6, “Using Fixed-Length

Format”
Examples of commands, format, and usage Appendix A, “Examples and

Grammar”
Command and Task Tips Appendix B, “Troubleshooting”
How to enter external table options on the com- Appendix C, “Option Names”
mand line, in a control file, or in a SQL command
Legal notices and trademark information Appendix D, “Notices and

Trademarks”
Symbols and Conventions

This guide uses the following typographical conventions:
 Italics for emphasis on terms and user-defined values such as user input
 Upper case for SQL commands; for example INSERT, DELETE
 Bold for command line input; for example, nzsystem stop
ix
If You Need Help
If you are having trouble using the IBM Netezza appliance, you should:
1. Retry the action, carefully following the instructions given for that task in the
documentation.
2. Go to the IBM Support Portal at: http://www.ibm.com/support. Log in using your IBM
ID and password. You can search the Support Portal for solutions. To submit a support
request, click the Service Requests & PMRs tab.
3. If you have an active service contract maintenance agreement with IBM, you can con-
tact customer support teams via telephone. For individual countries, visit the Technical
Support section of the IBM Directory of worldwide contacts (http://www14.soft-
ware.ibm.com/webapp/set2/sas/f/handbook/contacts.html#phone).
Comments on the Documentation

We welcome any questions, comments, or suggestions that you have for the IBM Netezza
documentation. Please send us an e-mail message at netezza-doc@wwpdl.vnet.ibm.com
and include the following information:
 The name and version of the manual that you are using
 Any comments that you have about the manual
 Your name, address, and phone number
We appreciate your comments on the documentation.
x
CHAPTER 1
Overview
What’s in this chapter
 Data Loading Components
 Data Loading Formats
 New Decimal Delimiter Option
This chapter provides general information about the data loading methods now available.
Note that loading data takes a significant allocation of system resources, which may affect
performance.
Data Loading Components

Within the IBM® Netezza® environment, data loading means simply to transfer data to the
IBM® Netezza® appliance. Within this framework, there are a number of components:
 External Tables – These are tables stored as flat files on the host or client systems and
not in the Netezza appliance database. These tables can be used to load data into the
Netezza appliance. For more information, see Chapter 2, “External Tables.”
 nzload – This is a command that provides an easy method for using external tables and
getting data into the Netezza appliance. For more information, see Chapter 4, “Using
nzload.”
 Format Options – These are options for formatting the data load to and from external
tables. Since data comes in different forms, Netezza provides different ways of setting
up the load. For more information, see Chapter 4, “Using nzload,” and Chapter 6,
“Using Fixed-Length Format.”
 Backup and Restore – There are different methods for doing backups and restores to
transfer data between systems. One method is to create external tables and use nzload,
described in Chapter 2, “External Tables,” and Chapter 4, “Using nzload.” For more
information on backups and restores, see “Backing Up and Restoring Databases” in
the IBM Netezza System Administrator’s Guide.
 nz_migrate – This is a separate tool, not part of the Netezza software package. This
utility is a script that can migrate (copy) a database/table from one Netezza appliance
to another, or make a copy of a database/table on the same server. Run the following
command to see the help explanation text for the command, showing syntax and usage:
nz_migrate -?
1-1
Netezza Data Loading Guide
Data Loading Formats

In the database environment, there is always the need to load data from external sources
such as files, pipes, or sockets into a table. These external sources have a variety of formats
to represent each of the data types individually, and together as records or rows.
When you load data from database-like applications, such as an RDBMS, a Web-server, or
some other structured data-store, they may export data into files or streams in different for-
mats. The following formats are used with the Netezza environment:
 Text-Delimited – The method commonly used for data loading is Text-Delimited format,
where every value of a field or column ends with a delimiter, and each set of these val-
ues of rows or records has an end-of-record delimiter, typically a new-line character.
Previously, this has been the preferred method used for loading data into external
tables.
 Fixed-Length – The new format, which allows for a more expressive form of external
table definition, thus increasing the kinds of data formats and layouts that can be
loaded.
 Compressed Binary – This Netezza proprietary format compresses the data before a
backup or restore to benefit performance. It typically yields smaller data files, retains
information about the Netezza appliance topology, and thus is often faster to backup
and restore. Compress the data before loading, and uncompress before unloading. For
more information, see the IBM Netezza System Administrator’s Guide.
New Decimal Delimiter Option

In the 6.0 release, a new option allows you to specify a comma as a decimal separator, in
addition to the period (the default value). This new option is available for external tables
and for nzload, to help you to directly load data without extra pre-load conversion.
 For the text-delimited format, and for unloading data, this option is available only at
the table level.
 For the fixed-length format, you can specify this option at the column level, making it
possible to have a mix of comma and decimal separators.
The option is available for the following data types, for both text-delimited and fixed-length
formats:
 Numeric
 Float
 Double
 Time
 Timetz
 Timestamp
Option usage for each data type is explained in each particular section describing that data
type. For examples of how to use this new option, see Appendix A, “Examples and
Grammar.”
1-2 20525 Rev.3

CHAPTER 2
External Tables
 About External Tables
 Command Syntax
 Transient External Tables
 Supported Data Types
 Restrictions
 Best Practices
 Examples
This chapter describes external tables, as well as best practices and restrictions for using
them. For options for using external table, see Chapter 3, “External Table Options.” For
examples of how to use external tables, see Appendix A, “Examples and Grammar.”
In the Netezza environment, there are the following types of tables:
 System tables – Stored on the host
 User tables – Stored on the SPUs
 External tables – Stored as flat files on the host or client systems
About External Tables

An external table allows Netezza to treat an external file as a database table. An external
table has a definition (a table schema), but the actual data exists outside of the Netezza
appliance database. External tables can be used to access files which are stored on the
Netezza host server or, in the case of a remote external table, Netezza can treat a file on a
client system as an external table (see REMOTESOURCE option).
After you have created the external table definition, you can use INSERT INTO statements
to load data from the external file into a database table, or SELECT FROM statements to
query the external table.
2-1
Privileges Required
To create an external table, you must have LIST privilege on the database and CREATE
EXTERNAL TABLE administration privilege. The database user who issues the CREATE
EXTERNAL TABLE command owns the resultant table. The operating system user must
have proper permission on the data object (READ permission for loading, WRITE permis-
sion for unloading).
Displaying External Table Information

To display information about external tables, use the \d command from the nzsql prompt.
 To list all external tables found in the current database, use the \dx command. For
example:
dev(admin)=> \dx
List of relations
Name | Type | Owner
-------------+-----------+-------
extlineitem | ext table | admin
xlineitem | ext table | admin
(2 rows)
 To list the options defined in an external table, use the \d <external_tablename> com-
mand. For example:
dev(admin)=>\d extlineitem
Log Files
By default, loading errors go into the following log files:
 nzbad – <tablename>.<dbname>.nzbad
 nzlog – <tablename>.<dbname>.nzlog
You can override the default by specifying a file for errors to go by using the following with
a filename:
 bf <filename> for nzbad
 lf <filename> for nzlog
Usage
Use external tables to do the following:
 Load data into the Netezza appliance from an external table and structure the loading
operation to manipulate the data by using casts, joins, dropping columns, and so on.
 Store data outside the Netezza appliance, either to transfer to another application, or
as a table backup. See “Backing Up and Restoring” on page 2-4.
 Create an external table and use data from an external table as part of a SQL query.
The power of external tables is that the entire Extraction-Transformation-Loading (ETL) pro-
cess is mapped to plain SQL. Since a SQL-based ETL process can be initiated/executed
from any SQL client that can talk to the Netezza appliance, it reduces or avoids the require-
ment of specialized ETL tools.
2-2 20525 Rev.3

About External Tables
To load an external data file into the Netezza appliance as an external table, you can do
either of the following:
 Use a FROM clause of a SELECT SQL statement/command, like any normal table.
 Use a WHERE clause of an UPDATE or DELETE SQL statement.
To unload an external table into an external data file, use the table as the target table in
any of the following SQL statements:
 INSERT SQL
 SELECT INTO SQL
 CREATE TABLE AS SELECT SQL
All references to columns in the external table can be complex SQL expressions used for
the transformation of external data during a load/unload process. For more information, see
“Restrictions” on page 2-13.
Parsing
For loads, the sequence of rows are parsed one-by-one from the external data file, and con-
verted into internal records of the external table. There could be errors during the parsing of
each row, or each column. For example, there could be errors in identifying the column
value itself, as in the case of a missing delimiter. Or there could be errors during the con-
version from external format to internal records of the external table, such as alphabets
mentioned for an integer column in Text-Delimited format.
Each error is logged in detail in an nzlog file, and bad rows are logged in an nzbad file.
These files help user to identify bad rows in the external data file and correct them for
reloading. Depending on the load options of the external table in use, each bad row would
either cause the row to be skipped, or the entire load to be aborted. Similarly, each bad col-
umn of a bad row could cause the rest of the row to be ignored, or if possible to recover, the
load could continue to parse subsequent columns of the same row.
Note that if there is an error in the project-expression on the external table columns, then
the entire load is aborted and the transaction rolled back. Errors of this nature are not
logged in nzbad or nzlog files, as they are outside of the scope of the external table load
mechanism. Once the processing reaches the normal SQL engine, the external table is
treated as if it is a normal table.
Unlike an external table that has external rows in an ordered sequence, normal user tables
have no implicit row order other than hidden rowid columns. So there is no way for a user
not using rowids to identify the bad row in a SQL engine. In this case, the Netezza system
just returns an error that a particular column caused an error, without identifying the bad
row. It is as if the query was selecting from a normal table and inserting into another nor-
mal table, with some row that caused the error during insertion.
20525 Rev.3 2-3

Backing Up and Restoring

You can use external tables to back up a table in the system database. While the Netezza
appliance database backup utility, nzbackup, enables you to create backups of the entire
database, the external table backup method allows you to create a backup of a single table,
with the ability to later restore it to the database as needed.
 To back up table data using an external table, create external table definitions for each
user table and then use SQL to insert into the external table.
 When you restore table data, create a table definition (if it does not exist) and then use
SQL to insert into the table from an external table.
Command Syntax
The CREATE EXTERNAL TABLE command has the following syntax.
 To create an external table based on another table:
CREATE EXTERNAL TABLE table_name
SAMEAS table_name
USING external_table_options
 To create an external table by defining columns:
CREATE EXTERNAL TABLE table_name
({ column_name type
[ column_constraint [ ... ] ]} [, ... ]
)
[USING external_table_options]
Note: The system permits and maintains PRIMARY KEY, DEFAULT, UNIQUE, and REFER-
ENCES. UNIQUE, PRIMARY KEY and REFERENCES are ignored for external tables. The
system does not support constraint checks and referential integrity. The user must ensure
constraint checks and referential integrity. For more information, see “Column Constraint
Rules for Empty Strings” on page 2-10.
Transient External Tables

Transient external tables (TET) provide a way to define an external table that exists only for
the duration of a single query. Transient external tables have the same capabilities and lim-
itations as normal external tables. A special feature of a TET is that the schema does not
have to be defined when the TET is used to load data into a table or when the TET is cre-
ated as the target of a SELECT statement.
Syntax
The following is the syntax for a TET:
INSERT INTO <table> SELECT <column_list | *>
FROM EXTERNAL 'filename' [(schema_definition)]
[USING (external_table_options)];
CREATE EXTERNAL TABLE 'filename' [USING (external_table_options)]

AS select_statement;
2-4 20525 Rev.3

Transient External Tables
SELECT <column_list | *> FROM EXTERNAL 'filename' (schema_definition)

[USING (external_table_options)];
Explicit Schema Definition

The schema of a transient external table can be explicitly defined in a query. When defined
this way, the schema definition is the same as is used when defining a schema using CRE-
ATE TABLE.
SELECT x, y, NVL(dt, current_date) AS dt FROM EXTERNAL '/tmp/test.txt'
( x integer, y numeric(18,4), dt date ) USING (DELIM ',');
The explicit schema definition feature can be used to specify fixed length formats.
SELECT * FROM EXTERNAL '/tmp/fixed.txt' ( x integer, y numeric(18,4),
dt date ) USING (FORMAT 'fixed' LAYOUT (bytes 4, bytes 20, bytes 10));
The SAMEAS keyword can also be used to specify that the schema of the external table is
identical to some other table that currently exists in the database.
SELECT * FROM EXTERNAL '/tmp/test.txt' SAMEAS test_table
USING (DELIM ',');
Implicit Schema Definition

If the schema is not explicitly defined, the schema for a transient external table is deter-
mined based on the query being executed. When a TET is used as a data source for an
INSERT statement, the external table will take on the schema of the target table.
The external table in this INSERT statement takes on the schema of the target table. The
columns in the external data file must be in the same order as the target table, and every
column in the target table must also exist in the external table data file.
INSERT INTO target SELECT * FROM external '/tmp/data.txt'
USING (DELIM '|');
Exporting Data Using Transient External Tables

A transient external table can also be used to export data out of the database. In this case
the schema of the external table is based on the query being executed.
Example:
CREATE EXTERNAL TABLE '/tmp/export.csv' USING (DELIM ',') AS
SELECT foo.x, bar.y, bar.dt FROM foo, bar WHERE foo.x = bar.x;
Remote Transient External Tables

A session connected to Netezza using ODBC, JDBC, or OLE DB from a client system can
import and export data using a remote transient external table, which is defined by using
the REMOTESOURCE option in the USING clause.
For example, the following SQL statement loads data from a file on a Windows system into
a TEMP table on Netezza, using an ODBC connection.
CREATE TEMP TABLE mydata AS SELECT cust_id, upper(cust_name) as name
from external 'c:\customer\data.csv' (cust_id integer, cust_name
varchar(100)) USING (DELIM ',' REMOTESOURCE 'ODBC');
20525 Rev.3 2-5

Remote external table loads work by sending the contents of a file from the client system to
the Netezza server where the data is then parsed. This method minimizes CPU usage on
the client system during a remote external table load.
Supported Data Types

Table 2-1 describes the Netezza supported data types for external tables. For more infor-
mation about the specific data types, see the IBM Netezza Database User’s Guide.
Table 2-1: Supported Data Types
Data Type Example Description
byteint 120 See “Integer Data Types” on page 2-7.

smallint 0
integer 256
bigint 1290985
numeric -99.56 See “Fixed-Point Data Types” on

decimal 123.679 page 2-7.
real –81293.35 See “Floating-Point Data Types” on

double precision page 2-8.
char (n) salary See “Character Strings” on page 2-10

and “Column Constraint Rules for Empty
Strings” on page 2-10.
varchar (n) I am John See “Character Strings” on page 2-10

and “Column Constraint Rules for Empty
Strings” on page 2-10.
boolean true An ASCII string that contains any of the

following values:
[true|false]|[yes|no]|[1|0]|[t|f]|[y|n]
See “BoolStyle” on page 3-3.
date 2002-02-04 The date is an exact four-byte data type.

The system recognizes a range of dates
composed of year, month, and day. See
“DateStyle” on page 3-5.
time 01:59:45 See “Time” on page 2-11.

23:00:01
time with time zone 01:15:33 -05 See “Time with time zone” on
page 2-12.
timestamp 2002-02-04 01:15:33 See “Timestamp” on page 2-12.
2-6 20525 Rev.3

Integer Data Types

Integer types are exact data types. The system generates an error if an input field’s value
cannot be expressed without loss of accuracy in the target table.
Table 2-2 describes the integer syntax.
Table 2-2: Integer Description
Syntax [‘+’|’-’]<digit>…
Description • Optional leading sign

• Unlimited leading zeros
• At least one decimal digit
Limitation • No thousands-separator commas

• No support for exponential notation
Table 2-3 describes the integer handling.
Table 2-3: Integer Handling
SQL Alias Representation Values
byteint int1 1 byte, signed min value = -128

max value = 127
smallint int2 2 bytes, signed min value = -32768

max value = 32767
integer int or int4 4 bytes, signed min value = –2147483648

max value = 2147483647
bigint int8 8 bytes, signed min value = –9223372036854775808

max value = 9223372036854775807
Fixed-Point Data Types

The fixed-point data types are exact data types. The system generates an error if an input
field’s value cannot be expressed without loss of accuracy in the target table or database.
Table 2-4 lists and describes the fixed-point syntax.
Table 2-4: Fixed-Point Description
Syntax [‘+’|’-’]<digit>…[‘.’[<digit>…]]
[‘+’|’-’]’.’<digit>…
[‘+’|’-’]<digit>…[‘,’[<digit>…]]
[‘+’|’-’]’,’<digit>…

20525 Rev.3 2-7

Table 2-4: Fixed-Point Description
Limitation • No thousands-separator commas

• No support for exponential notation
The syntax of fixed-point values is the same as the syntax of integer values with the addi-
tion of an optional decimal digit that can occur anywhere — from before the first decimal
digit to after the last decimal digit.
The optional decimal point can be followed by zero or more decimal digits, if there is at
least one decimal digit before the decimal point; followed by one or more decimal digits if
there are no decimal digits before the decimal point.
If there is no explicit decimal point, the system assumes a decimal point immediately fol-
lowing the last decimal digit.
You can also specify a comma as a separator, using it like the decimal digit. For examples
of how to do this, see “Decimal Delimiter Examples” on page A-4.
Table 2-5 describes the fixed-point precision and representation:
Table 2-5: Fixed-Point Precision
Precision P≤9 9 < P ≤ 18 18 < P ≤ 38
Representation 4 bytes, signed 8 bytes, signed 16 bytes signed
The following result in system errors:

 Precision – Having more decimal digits before the decimal point than the declaration
allows (P-S).
 Scale – Having more decimal digits following the decimal point than the declared scale
(S).
Note: Because fixed-point is an exact data type, when there are too many digits following
the decimal point, the system does not round the number.
Floating-Point Data Types

The floating-point data types are approximate data types. The system rounds the signifi-
cand if more precision is present that it can represent.
Table 2-6 lists the floating point syntax.
Table 2-6: Floating-Point Description
Syntax [ '+' | '-' ] <digit>… [ '.' [ <digit>… ] ] [( 'e' | 'E' ) [ '+' | '-' ] <digit>… ]
[ '+' | '-' ] '.' <digit>… [ ( 'e' | 'E' ) [ '+' | '-' ] <digit>… ]
[ '+' | '-' ] <digit>… [ ',' [ <digit>… ] ] [( 'e' | 'E' ) [ '+' | '-' ] <digit>… ]
[ '+' | '-' ] ',' <digit>… [ ( 'e' | 'E' ) [ '+' | '-' ] <digit>… ]
2-8 20525 Rev.3

Table 2-6: Floating-Point Description

• Decimal point or comma, if needed
• Optional ‘e’ or ‘E’ introducing an exponent followed by an optional sign
and one or more digits
Limitation • No thousands commas

• No support for loading exceptional values (Not a Number (NaNs) and
infinities)
The syntax of fixed-point values is the same as the syntax of fixed-point values augmented
by an optional trailing exponent specification.
The optional decimal point can be followed by zero or more decimal digits, if there is at
least one decimal digit before the decimal point; followed by one or more decimal digits if
there are no decimal digits before the decimal point.
If there is no explicit decimal point, the system assumes a decimal point immediately fol-
lowing the last decimal digit.
You can also specify a comma as a separator, using it like the decimal digit. For examples
of how to do this, see “Decimal Delimiter Examples” on page A-4.
The optional power of ten exponent is ‘e’ (lower or uppercase), with an optional sign,
non-empty sequence of decimal digits.
Table 2-7 describes the floating-point precision and representation:
Table 2-7: Floating-Point Precision
Type Real Double
Representation 4 byte IEEE floating 8 byte IEEE floating

point point
Approx. largest normalized value ±3.40e+38 ±1.79e+308
Approx. smallest normalized value ±1.18e-38 ±3.40e-308
Approx. smallest denormalized ±7.01e-46 ±2.50e-324

value
The following result in system errors:

 Overflow – If the field exceeds the largest representable value (maximal exponent and
maximal significand)
 Underflow – If the number is too small to approximate in the denormalized range
20525 Rev.3 2-9

Character Strings
Char(n)/nchar(n) are character strings of length n. Varchar(n)/nvarchar(n) are vari-
able-length character strings of maximum length n. A valid character is between the ASCII
values 32 to 255.
System Handling of Characters

Table 2-8 describes how the system handles char, nchar, varchar, and nvarchar characters.
Table 2-8: Character Handling
Char, Nchar, Varchar,

How Handled
and Nvarchar
Padding Char/Nchar – Padded to normal length with spaces

Varchar/Nvarchar – Not padded
Truncation If the data is longer than the field:

• The system writes the record to the nzbad file.
• The system writes the records and column number to the nzlog
file.
Note that you can turn on automatic truncation with the -trunc-
String option.
Note: If you use this option for Unicode character data, it could
truncate combined NFC characters if they exceed the specified
column length. The switch does not attempt to keep any grapheme
clusters; it truncates data as necessary to fit in the specified
column size.
Column Constraint Rules for Empty Strings

For all char(n) and varchar(n) data types, the result of inserting an empty string and filling
in missing data values depends upon whether the columns are declared null-able (default)
or not null-able (declared with constraint not null). Table 2-9 describes the different cases.
Table 2-9: Column Constraining Rule for Empty Strings
Column Null Token

Data Type Null Token Exists
Constraint Does Not Exist
null token "" (empty string) "" (empty string)
Char/Nchar NULL NULL char/nchar: space NULL

Varchar/Nvarchar filled.
varchar/nvarchar: zero
length string.
NOT NULL ERROR char/nchar:: space ERROR

filled.
varchar/nvarchar: zero
length string.
2-10 20525 Rev.3

Table 2-9: Column Constraining Rule for Empty Strings
Column Null Token

Data Type Null Token Exists
Constraint Does Not Exist
null token "" (empty string) "" (empty string)
Bool, Date, NULL NULL NULL NULL

Int (1,2,4,8),
Numeric(), NOT NULL ERROR ERROR ERROR
Float (4,8), Time,
Timestamp, Timetz
If the record contains fewer data values than the actual columns defined in the table’s
schema, the system writes an error to the nzlog file and discards the record. To override this
behavior, use the -fillRecord option, which applies to the entire load operation.
The -fillRecord option tells the system to use a null value in place of any missing fields.
You can use this option as long as the columns whose values are missing allow nulls. If
these columns are defined as not null, the system writes an error to the nzlog file and dis-
cards the record. You must resolve this conflict by changing the schema to allow null values
or modifying the data file to include a valid non-null value.
Time Data Types

The system supports time, timestamp, and time with time zone. These data types are exact
types, stored to the accuracy of 1 μ Sec (1/1,000,000 of a second).
You can also specify a comma as a separator in time data types, using it like the decimal
digit. For examples of how to do this, see “Decimal Delimiter Examples” on page A-4.
Time
The Netezza appliance time is an exact, eight-byte data type stored internally as a signed
integer representing the number of microseconds since midnight.
The system accepts both 24 hour and 12 hour AM/PM time values. You can specify the for-
mat with the -timeStyle option. The default is the 24-hour format.
The time format consists of five components: hour, minute, second, fraction of a second,
and AM/PM token. You must have hour and minute; second and fraction of second are
optional. The AM or PM token is required for 12 hour and not allowed for 24-hour format.
The time options have the following formats. Note that the delimited examples use the
default time delimiter, which is a colon (:).
 12-hour delimited HH:MM:SS.FFF [AM | PM] (such as 10:12 PM, or 1:02:46.12345
AM)
 12-hour undelimited HHMMSS.FFF [AM | PM] (such as 1012 PM or 010246.12345
PM)
 24-hour delimited HH:MM:SS.FFF (such as 19:15 or 1:15:00.1234)
 24-hour undelimited HHMMSS.FFF (such as 1915 or 10246.12345 PM)
In these formats, note the following:
20525 Rev.3 2-11

 HH is a one- or two-digit hour value from 1–12 in the 12-hour notation or 1–24 in the
24-hour notation. In undelimited format, you must specify two digits such as 01, 02,
and so on.
 MM is a one- or two-digit minute value from 1–60. In undelimited format, you must
specify two digits such as 01, 02, and so on.
 SS is a one- or two-digit seconds value from 1–60. In undelimited format, you must
specify two digits such as 01, 02, and so on.
 FFF specifies a fraction of a second. If you specify a fractional value, you must precede
it with a decimal point. If the value can be stored without loss of precision, it is
accepted. If the value cannot be stored without loss of precision, it is rejected. You can
use the -timeRoundNanos option to allow rounding when the full precision of any frac-
tional digits cannot be preserved, as described in “Using the -timeRoundNanos
Option” on page 10-22.
Time with time zone

The Netezza time with time zone (timetz) is an exact data type stored in 12 bytes. Inter-
nally the Netezza appliance stores it as time and an offset. The stored offset has the same
1μS resolution as time even though the input is restricted to a one-minute resolution.
Syntax The input format of time with time zone value is identical to that of simple time
followed by a trailing signed offset from Coordinated Universal Time (UTC — formerly
Greenwich Mean Time GMT). The time section must conform to the -timeStyle and
-timeDelim in effect during the nzload job.
You must specify a signed, time-zone hour, whereas the time-zone minute is optional. If
you use the minute, separate it with a colon (the default timeDelim character).
Note: You cannot use named time zones, such as EST.
Table 2-10 lists the time with time zone syntax.
Table 2-10: Time With Time Zone Description
time with time zone <time> ( '+' | '-' ) <digit> [ <digit> [ ':' <digit> [ <digit> ] ] ]
Errors The following are time and range errors:

 Time – The same errors as the time data type.
 Range – The time zone offset is restricted to -13:00 to +12:59.
Timestamp
The Netezza appliance timestamp is an exact data type stored as eight bytes. The stored
offset has the same 1μS resolution as the time data type.
Syntax The input format of a timestamp value is a date value followed by a time value.
You can have optional spaces between the date and the time. The date section must con-
form to the -dateStyle and -dateDelim in effect during the load job.
2-12 20525 Rev.3

Restrictions
Table 2-11 lists the timestamp syntax.
Table 2-11: Timestamp Description
timestamp <date> <time>
Errors The following are date and time errors:

 Date – The same errors as the date data type.
 Time – The same errors as the time data type.
Restrictions
The following are restrictions and considerations for use with external tables:
 Always consider your source and target systems, and whether the data is properly for-
matted for loading.
 To insert and drop an external table, use the INSERT and DROP commands.
 You cannot delete, truncate, or update an external table. After creating an external
table, you can alter as well as drop the table definition. (Dropping an external table
deletes the table definition, but it does not delete the data file that is associated with
the table.) You can select the rows in the table, as well as insert rows into the table
(following a table truncation).
 While you cannot select from more than one external table at a time in a query or sub-
query, you can move data from one external table to another, such as using SELECT
and INSERT. The system displays an error if you incorrectly specify multiple external
tables in a SQL query, or if you reference the same external table more than once in a
query:
*ERROR: Multiple external table references in a query not
allowed*
To specify more than two external tables, load the data in into a non-external table and
specify this table in the query.
 You cannot perform a union operation involving two or more external tables.
 Using the nzbackup command to back up external tables backs up the schema but not
the data.
 Host-side operations, such as selects and rowsetlimit user and group property interac-
tions, are not supported for compressed external tables.
 The DecimalDelim option is not supported for compressed external tables.
 There is a maximum limit of 300 concurrent loads for multiple loads.
Best Practices
When specifying external tables, note the following:
20525 Rev.3 2-13

 An external table reference can appear as the source table of a SELECT FROM state-
ment. Note that a transient external table reference in a SELECT FROM clause infers
its shape from the preceding INSERT INTO clause.
 In Netezza Release 4.6 and later, the system catalog datatypes TEXT and NAME are
treated as NVARCHAR. If these types are used in the table that is referenced in the
select_clause, include the encoding option in the CREATE EXTERNAL TABLE com-
mand to specify internal encoding. Otherwise you could receive the error “LATIN9
encoding cannot be specified with NCHAR/NVARCHAR column definitions.” For
example:
create external table '/tmp/ext1' using (encoding 'internal') as
select username from _t_user;
 The CREATE EXTERNAL TABLE AS statement supports an optional table name. If you
do not provide a table name, the table is transient, which means the external table def-
inition does not persist in the system catalog. If you supply a table name, the external
table becomes a named object in the system catalog.
 The USING clause in the inline external statement is optional. If you omit it, the result-
ing external table has the default settings. Note that you must specify the USING
clause in the CREATE EXTERNAL TABLE SAMEAS statement, because the SAMEAS
table might be another external table.
 When you insert data into an external table that references an existing data file, the
system truncates the file before inserting the external table’s data.
 You cannot use external tables in complex SQL statements. If the statement is not sup-
ported, the system displays an error.
Before you reload an external table, verify that the destination table in the database is
empty or that it does not already contain the rows in the external table that you are about to
reload. If the destination table already contains the rows contained in the external table,
unintended problems may occur. These problems could also occur if you accidentally
reload the external table more than once.
For example, loading a text-format external table into a destination table that already con-
tains the same data creates duplicate data in the database. The rows will have unique row
IDs, but the data will be duplicated. To fix this problem, you would have to delete the
duplicate rows or truncate the database table and reload the external table again (but only
once).
If you load a compressed binary format external table into a destination table that already
has the same rows, you will create duplicate rows with duplicate row IDs in the database
table. The system restores the rows using the same row IDs saved in the compressed binary
format file.
Duplicate row IDs can cause incorrect query results and could lead to problems in the data-
base. You can check for duplicate rowIDs using the rowid keyword as follows:
SELECT rowid FROM employee_table GROUP BY rowid HAVING count(rowid)
>1;
If the query returns multiple rows that share the same row ID, truncate the database table
and reload the external table (but only once).
2-14 20525 Rev.3

Examples
After you load data from an external table into a user table, you should run GENERATE
STATISTICS to update the statistics for the user table. This improves the performance of
queries that run against that table.
Examples
The following examples show how to use the CREATE EXTERNAL TABLE command.
 To create an external table, enter:
CREATE EXTERNAL TABLE ext_orders(ord_num INT, ord_dt
TIMESTAMP)USING(dataobject('/tmp/order.tbl') DELIMITER '|');
 To create an external table that uses column definitions from an existing table, enter:
CREATE EXTERNAL TABLE demo_ext SAMEAS emp USING (dataobject
('/tmp/demo.out') DELIMITER '|');
 To create an external table and specify the escape character (‘\’), enter:
CREATE EXTERNAL TABLE extemp SAMEAS emp USING( dataobject
('/tmp/extemp.dat') DELIMITER '|' escapechar '\');
 To unload data from your database into a file by using an insert statement, enter:
INSERT INTO demo_ext SELECT * FROM emp;
 To drop an external table, enter:
DROP TABLE extemp
The system removes only the external table’s schema information from the system cat-
alog. The file defined in the dataobject option remains unaffected in the filesystem.
 To back up by creating an external table, enter:
CREATE EXTERNAL TABLE '/path/extfile' USING (FORMAT 'internal'
COMPRESS true) AS SELECT * FROM source_table;
 To restore from an external table, enter:
INSERT INTO t_desttbl SELECT * FROM EXTERNAL'/path/extfile'
USING(FORMAT 'internal' COMPRESS true);
Transient External Table

The following examples show how to specify the shape of a transient external table:
 To take on the schema of the target table:
insert into <table> select * from external '<file>' [USING(...)]
 To take on the schema of the query:
create external table '<file>' [USING (...)] as <QUERY>
 To take on the schema of <table>:
select * from external '<file>' sameas <table> [USING(...)]
 To take on the schema as defined:
select * from external '<file>' (schema) [USING(...)]
 To take on the schema as defined:
create external table '<file>' (schema) [USING(...)]
20525 Rev.3 2-15

 To make the source file FIXED format with the schema as defined:
select * from external '<file>' (schema) USING (FORMAT 'FIXED'
LAYOUT (...))
 To make the source file FIXED format and the table takes on the schema of the target
table:
insert into <table> select * from external '<file>' USING (FORMAT
'FIXED' LAYOUT (...))
 The following example will not work, because you cannot unload data into a FIXED for-
mat external table:
create external table '<file>' [(schema)] USING (FORMAT 'FIXED'
LAYOUT ... )
Fixed-Length Format
The following examples show how to use Fixed-Length format with external tables:
 To load data in fixed format, enter:
INSERT INTO t SELECT * FROM EXTERNAL ‘/data/fixed’ USING ( FORMAT
‘FIXED’ LAYOUT (BYTES 20, REF BYTES 3, BYTES @2) )
 To load data with different date/time delimiters for different zones, enter:
‘FIXED’ LAYOUT ( YMD ‘-‘ BYTES 15, DMY ‘/’ BYTES 15 ) )
 To load spatial data (binary data into VARCHAR), enter:
‘FIXED’ CTRLCHARS true LAYOUT ( BYTES 100, REF BYTES 4, BYTES @2) )
 To load fixed format data with record-length and no record-delimiter, enter:
‘FIXED’ RECORDDELIM ‘’ RECORDLENGTH @1 LAYOUT( REF BYTES 2, BYTES
120, REF BYTES 2, BYTES @3) )
 To load data with different NULLIF clauses for different zones, enter:
‘FIXED’ LAYOUT ( BYTES 15 NULLIF ‘2000-10-10’, BYTES 2 & = ‘12’) )
 To load data with NULLIF clauses referring other zones, enter:
‘FIXED’ LAYOUT ( REF BYTES 2, BYTES @1 NULLIF @1 = -1, REF BYTES 4,
BYTES 100 NULLIF &&3 = ‘null’ ) )
Standard Unloading and Reloading

The following examples unload and load a user table to an external table in text-delimited
format. Unloading is not supported for Fixed-Length format.
 To create a text-format external table, enter:
CREATE EXTERNAL TABLE extemp SAMEAS emp USING (DATAOBJECT
('/tmp/emp.dat'));
 To unload data in user table EMP to the external table EXTEMP, enter:
INSERT INTO extemp SELECT * FROM emp;
2-16 20525 Rev.3

Examples
 To load data into user table EMP from external table EXTEMP, enter:
TRUNCATE TABLE emp;
INSERT INTO emp SELECT * FROM extemp;
Back up and Restore a User Table

The following examples show how to back up and restore the user table EMP to an external
table in binary compressed format.
 To create a compressed binary format external table definition called emp_backup for
the table emp, enter:
CREATE EXTERNAL TABLE emp_backup SAMEAS emp USING (
DATAOBJECT ('/tmp/emp.bck')
COMPRESS true
FORMAT 'internal');
 To back up the emp table data into emp_backup, enter:
INSERT INTO emp_backup SELECT * FROM emp;
 To restore the emp table from emp_backup, make sure that the emp table is empty and
enter:
TRUNCATE TABLE emp;
INSERT INTO emp SELECT * FROM emp_backup;
20525 Rev.3 2-17

2-18 20525 Rev.3

CHAPTER 3
External Table Options
 Options
 Option Details
 Option Processing
 Session Variables
This chapter describes the options used with external tables. For examples of how to use
external tables, see Appendix A, “Examples and Grammar.”
Options
When you create an external table definition, you can specify options. There are different
types of options: some are for records/rows, some are for fields, and some are for loads. Use
these options when loading from an external table or when using the external table directly
in a SQL query.
Note: The best method to verify that the load processing has been successful is to ensure
the system records any errors to the nzlog and nzbad files. Check these files occasionally.
Table 3-1 lists the external table options, and a description of each option follows. In the
Valid Formats column, Text refers to Text-Delimited format and Fixed refers to Fixed-Length
format. In the Data type column, enumeration refers to the system accepting a specified
set of quoted or unquoted string values.
Table 3-1: External Table Options
Valid Unload
Option Values Default Data Type
Formats Y/N
BoolStyle Text, 1_0/T_F/Y_N… NULL, 1_0 Y enumera-

Fixed tion
Compress Text, True/False False Y boolean

Fixed
CRinString Text, True/False NULL, False Y boolean

Fixed
3-1
Valid Unload
Formats Y/N
CtrlChars Text, True/False NULL, False N boolean

Fixed
DataObject Text, Existing file path No default Y filename

Fixed
DateDelim Text, 1-byte NULL, "-" Y string

Fixed
DateStyle Text, YMD/MDY/DMY… NULL, YMD Y enumera-

Fixed tion
DecimalDelim Text, 1-byte ‘.’ Y string

Fixed
Delimiter Text 1-byte NULL, "|" Y string
Encoding Text Inter- NULL, Y enumera-

nal/Latin9/Utf8 Internal tion
EscapeChar Text 1-byte NULL Y string
FillRecord Text True/False NULL, False N boolean
Format Text, Text/Inter- Text Y enumera-

Fixed nal/Fixed tion
IgnoreZero Text True/False NULL, False N boolean
IncludeZero- Text True/False NULL, False Y boolean

Seconds
Layout Text, Zone definitions NULL, Inherit N none

Fixed
LogDir Text, existing dir path NULL, /tmp N string

Fixed
MaxErrors Text, >=0 NULL,1 N integer

Fixed
MaxRows Text, >=0 NULL, 0 N integer

Fixed
NullValue Text, 4-bytes NULL, Y string

Fixed "NULL"
QuotedValue Text No/Yes/Sin- NULL, No N enumera-

gle/Double tion
3-2 20525 Rev.3

Option Details
Valid Unload
Formats Y/N
RecordDelim Text, 4-bytes NULL, N string

Fixed /newline
RecordLength Fixed Integer/Zone-ref NULL N integer

expr
RemoteSource Text, ODBC/JDBC/OLE- NULL Y enumera-

Fixed DB tion
RequireQuotes Text True/False NULL, False N boolean
SkipRows Text, >=0 NULL, 0 N bigint

Fixed
SocketBufSize Text, 64KB-2GB 8MB Y integer

Fixed
TimeDelim Text, 1-byte NULL, ":" Y string

Fixed
TimeRound Text True/False NULL, False N boolean

Nanos
TimeExtraZeros
TimeStyle Text, 24hour/12hour NULL, Y enumera-

Fixed 24hour tion
TruncString Text True/False NULL, False N boolean
Y2Base Text, >=0 NULL, 0 N integer

Fixed
Option Details
The following sections details the different options.
BoolStyle
Specifies the boolean style. During a load, the loader can handle only a specific style of
boolean values.
Table 3-2 lists the styles and their values.
Table 3-2: Boolean Values
Style Name Value
1_0 1 or 0
20525 Rev.3 3-3

Table 3-2: Boolean Values
Style Name Value
T_F T or F
Y_N Y or N
YES_NO YES or NO
TRUE_FALSE TRUE or FALSE
The default style is 1_0. The values can be expressed in mixed case, so ‘true’ can be ‘True’
or ‘TRUE’ or ‘tRuE’.
If you specify the YES_NO option on the command line, the system assumes that the data
in the Boolean field is in the form yes or no. If the data is any of the other values: true,
false, 1, 0, t, f, y, or n, the system discards the record to the nzbad file and logs an error
with the record number in the nzlog file.
Compress
Specifies whether the source datafile data is compressed or not. The valid values are true
or on, false or off. The default is false. This can only be true if the format is set to ‘internal’.
CRinString
Specifies whether to allow unescaped carriage returns in char/varchar and nchar/nvarchar
fields. Acceptable values are true or false, on or off. Do not put quotes around the value.
 False – Default, treats all cr or crlf as end-of-record.
 True – Accepts unescaped CR in char/varchar fields (LF becomes only end of row).
Note: This option is different for Fixed-Length format. For more information, see “Changed
Options” on page 6-3.
CtrlChars
Specifies whether to allow an ASCII value 1-31 in char/varchar and nchar/nvarchar fields.
You must escape NULL, CR, and LF characters. Acceptable values are: true or false, on or
off. The default is false. Do not insert quotes around the value.
DataObject
Specifies the OS-path to the source datafile (or any media that can be treated as a file).
There is no default, and this must be specified. When the remotesource option is not set (or
set to empty string), this path has to be an absolute path and not a relative path. The file-
name must be a valid UTF-8 string.
 For loads, this file has to be an existing file with READ permission for the OS user ini-
tiating the load.
3-4 20525 Rev.3

Option Details
 For unloads, the parent directory of this file has to have READ-WRITE permissions for
the OS user initiating the unload, and the data file is overwritten if it already exists.
DateDelim
Specifies the delimiter character that separates the date components, and used with the
dateStyle option. The default is ‘-‘ for all dateStyles except MONDY[2], where the default is
‘ ‘ (space). This is a single-byte string.
 If you specify the option as an empty string, which means that there is no delimiter
between the date components, you must specify days and months as two-digit num-
bers. Single-digit months and days are not supported.
 With MonDY or MonDY2, the default dateDelim option is space.
 With days and months less than 10, use either one or two digits, or a space followed by
a single digit.
 With the dateDelim option as a space, the system allows a comma after the day.
 With any component (day, month, year) as zero, or any day/month inconsistency, such
as August 32 or February 30, the system returns an error.
Table 3-3 lists dateDelim option examples.
Table 3-3: The -dateDelim
No dateDelim -dateDelim ’,’ -dateDelim ’ ’ (space)
Jan 01 2003 Jan 01,2003 Jan 01, 2003
Jan 1 2003 Jan 1,2003 Jan 1, 2003
Jan 1 2003 Jan 1,2003 Jan 1, 2003
Note: If not using delimiters, the date will be determined as in the following example for
June 12, 2009:
06122009
DateStyle
Specifies how to interpret the date format. The date style settings ‘YMD’, ‘MDY’, ‘DMY’,
‘DMONY’, ‘MONDY’. The default is YMD.
Note: The two-digit year formats (Y2MD, MDY2, DMY2, DMONY2 and MONDY2) are not
supported for unloads.
The dateStyle options are shown in Table 3-4, with an example date of March 21, 2012.
The examples shown have no date delimiter between the values.
Table 3-4: DateStyle
Sequence of Date Components Four-digit Year Two-digit Year
Year Numeric-month Day YMD Y2MD

(20120321) (120321)
20525 Rev.3 3-5

Table 3-4: DateStyle
Sequence of Date Components Four-digit Year Two-digit Year
Day Numeric-month Year DMY DMY2

(21032012) (210312)
Numeric-month Day Year MDY MDY2

(03212012) (032112)
Alphabetic-month Day Year MonDY MonDY2

(Mar212012) (Mar2112)
Day Alpha- Year DMonY DMonY2

betic-month (21Mar2012) (21Mar12)
For example, if your datafile has a date of 03-21-12, with a DateDelim of ‘-’, use the for-
mat MDY2.
The default dateStyle is YMD, and the SQL standard stipulates that the legal years are
0001 to 9999. There is no provision in SQL for years prior to 0001 AD or later than 9999
AD.
Date example: In the data file jan-01.data, the data are specified as the following (with the
date format shown in Bold):
14255932|30/06/2002|20238|20127|40662|157|
Because the date value is using the DD/MM/YYYY format, specify the following dateStyle
and dateDelim options:
nzload -t agg_month -df jan-01.data -delim ‘|’ -dateStyle DMY -datedelim '/'
DecimalDelim
Specifies the decimal delimiter for the following data types, for both text-delimited and
fixed-length formats: float, double, numeric, time, timetz, and timestamp. Default is ‘.’.
For examples of usage, see “Decimal Delimiter Examples” on page A-4.
Delimiter
Specifies the field delimiter. The default is the pipe character ‘|’. You can specify charac-
ters in the 7-bit ASCII range using either a quoted value (for example: delimiter '|') or
by its unquoted decimal number (delimiter 124) . To specify a byte value above 127,
use the decimal number. This is a single-byte string.
Note: For nzload, the default is ‘\t’(tab).
Note: This option is not supported for Fixed-Length format.
The system processes an input row by identifying the successive fields within that row. A
single character field delimiter separates adjacent fields. The lack of a field delimiter
between fields is an error. You can use a trailing field delimiter following the last field in a
row (but it is not required).
You can specify the following delimiters:
 Numeric – 0xNN or NN where NN is a number for either hexadecimal or decimal.
3-6 20525 Rev.3

Option Details
 Control characters – ^A -^Z (low-order 5 bits) and ^a -^z (low-order 5 bits).

 Symbols – \b backspace (8), \t horizontal tab (9), \n line feed (10), \f form feed (12), \r
carriage return (13), \\ backslash, \’ single quote, \” double quote.
 Literal – Any character, such as c (the non-control character c).
To use a character other than a 7-bit-ASCII character as a delimiter, make sure that you
specify it as a decimal or hex number. Do not specify a character literal, which could result
in errors from encoding transformation. For example, to use the hex value 0xe9 as a delim-
iter (which is é in Latin9), use –d 0xe9 as the value. Do not use –d 'é'.
Although the system accepts alpha-numeric characters, to avoid ambiguity do not select a
delimiter that conflicts with the data in a field. Also if you use the dateDelim and timeDe-
lim options, select different delimiters for each type.
Note: When you are using the nzload wrapper you can enter escape characters on the com-
mand line, such a \b. If you use the CREATE EXTERNAL TABLE command, the only special
character you can specify is \t (“\t”).
Encoding
Specifies the encoding of the datafile for the character set. The default is ‘internal’.
You can also specify ‘utf8’ if the whole file is in UTF-8 encoding and has only nchar/nvar-
char data and no char/varchar data. Use ‘internal’ if the file could have both Latin-9 and
UTF-8 data– or either type– using char, varchar, nchar, or nvarchar data.
The system supports single-byte characters in Latin9 encoding, and Unicode data in the
multi-byte UTF-8 encoding. Use the encoding option to specify the type of data in the file.
The encoding option has three values:
 A value of ‘latin9’ indicates that the whole file is in Latin-9 char/varchar data and has
no nchar/nvarchar data. (If the file contains any nchar/nvarchar data, it will be rejected
by the load operation.)
 A value of ‘utf8’ indicates the whole file is in UTF-8 encoding and has only nchar/nvar-
char data and no char/varchar data. (If the file contains any char/varchar data, it will be
rejected by the load operation.)
 The value ‘internal’ indicates that the file could have either or both Latin-9 and UTF-8
data using any or all of the char, varchar, nchar, or nvarchar data types. As a best prac-
tice, use ‘internal’ if you are not certain of the data encoding.
For more information, see the “Using International Character Sets” chapter in the IBM
Netezza Database User’s Guide.
Use the nzconvert command to convert character encoding before loading with external
tables. For the command options and examples, refer to “Converting Legacy Formats” in
the IBM Netezza Database User’s Guide.
EscapeChar
Specifies the use of an escape character. The character immediately following the ‘\’ is
escaped. The only supported value is ‘\’, and the default is no escaping.
By default, the system expects fields to be delimited by a field-delimiter character or by an
end-of-row sequence. The system assumes all other characters are part of the field’s value.
20525 Rev.3 3-7

Although efficient, this representation has the drawback that string fields may not contain
instances of the field delimiters. In addition, one value typically becomes inexpressible
because you have used it to convey the absence of any value (that is, that column is null).
One solution is to use an escape character for the delimiter. For example, the following
command line demonstrates using the escapeChar option.
nzload -escapeChar ‘\’ -nullValue ‘NULL’ -delim ‘|’
 |NULL| – A null input field
 |\NULL| – A non-null input field containing the text NULL
 |\|| – A non-null input field containing the single character |
 |\\| – A non-null input field containing the single character \
FillRecord
Specifies whether to allow an input line with fewer columns than the table definition. Miss-
ing or trailing input fields should be treated as nulls if the columns are nullable. The
default is false.
The system expects one input field for every column in the target table’s schema, and
rejects a row with fewer fields. If you specify the fillRecord option, the system allows omit-
ting one or more trailing (rightmost) fields, as long as all corresponding columns can be
null.
Format
Specifies the data format of the source file to load and unload. The valid values are as
follows:
 ‘text’ (default) – Data in Text-Delimited format
 ‘fixed’– Data in new Fixed-Length format
 ‘internal’ – Data in compressed binary format (to use this, the compress option must be
set to true)
IgnoreZero
Specifies discarding byte value zero in char() and varchar() fields. The default is false. If
true, the command accepts binary value zeroes in input fields and discards them.
IncludeZeroSeconds
Specifies that “00” seconds values will be unloaded to the external table. For example, a
time value such as 12:34:00 or 12:34 will be unloaded to the external table in the format
12:34:00. The default is false.
Note: This option is not supported for Fixed-Length format, and is only for unloading.
3-8 20525 Rev.3

Option Details
Layout
Specifies the zone definitions.
Note: This option is used only with the Fixed-Length format. For more information, see
“New Options” on page 6-2.
LogDir
Specifies the directory to which nzlog and nzbad files are generated for loads. This is not
used for unloads. The default value is '/tmp'. Note that when doing remote loads from Win-
dows clients (through ODBC/JDBC), the default output directory is mapped to "C:\". The
directory name must be a valid UTF-8 string.
MaxErrors
Specifies the number of errors at which the system stops processing rows. If the count of
rejected rows reaches this threshold, the system immediately aborts and rolls back the
load.
The default value is 1. This default has the effect of committing a load only if it contains
no errors. A maxErrors value n (where n is greater than 1) allows the first n-1 row rejections
to be recoverable errors, not including the number of rows processed in the skipped row
range.
Use this option to specify a different value, from 0 (unlimited errors) up to 2,147,483,647
(the largest signed 32-bit integer).
MaxRows
Specifies to stop processing after this initial number of rows. Use a limit clause with the
select statement to limit loading data. The default is 0 (load all rows).
After processing a row (whether inserted, skipped or rejected), the system decides whether
to look for another input row:
 If you did not specify the maxRows option, the system attempts to locate the next input
row.
 If you specified the maxRows option and the input row counter is equal to the
maxRows count, the system ends the load and commits all inserted records, not
including the rows processed in the skipped row range. Otherwise, the system attempts
to locate the next input row.
NullValue
Specifies the string to use for the null value, with a maximum 4-byte UTF-8 string. The
default is ‘NULL’. You can specify a value such as a space (' ') or any string up to four char-
acters. Conceptually a field contains either a value or an indication that there is no value.
The system provides some flexibility in how you indicate that a field contains no value. For
more information about how the system handles nulls, see “Column Constraint Rules for
Empty Strings” on page 2-10.
The system determines a field’s type and whether it is null by inspecting the corresponding
column declaration:
20525 Rev.3 3-9

 If there is no value, the system sets the corresponding value in the candidate binary
record to null.
 If you declared the target column “not null,” then an absence of a value is an error.
 If a field does not indicate null, the system assumes it contains a value. The system
analyzes the contents of that field, converts its textual input representation to binary,
and sets the corresponding value in the candidate binary record to that value.
QuotedValue
Specifies whether data values are quoted or not. The default is false. Specify SINGLE or
YES to require single quotes or DOUBLE to require double quotation marks. You can pre-
cede the opening quote or follow the closing quote with spaces. You can use the actual
quote characters if you enclose them in double quotes. The system recognizes the end of
the field by a field-delimiter character or an end-of-row sequence.
The system recognizes a quoted value when the first non-space character is the quote char-
acter specified in the quotedValue option. If the first non-space character is not the
specified quote character, then the system handles it according to the normal rules. In par-
ticular, leading or trailing spaces in string fields are considered part of the string’s value.
For example, the following command line demonstrates using the quotedValue option.
nzload -quotedValue SINGLE -nullValue ‘NULL’ -delim ‘|’
 |NULL| – A null input field
 |’NULL’| – A null input field
 | I’m | – A non-null input field containing the text “I’m “
 | ‘I’’m’ | – A non-null input field containing the text “I’m“
 | ‘|’ | – A non-null input filed containing the single character “|”
 |’ ‘| – A non-null input field containing a single space
 | | – A non-null input field containing a single space
 | ‘‘ | – A non-null input field containing a zero-length string
 || – A non-null input field containing a zero-length string
Note that unlike the escapeChar option, the quotedValue option is not able to force the sys-
tem to accept the nullValue token as a valid non-null input value. The system overhead for
processing quoted value syntax is much greater than the default unquoted syntax. In addi-
tion, except for strings containing three or more field delimiters that need to be escaped
and no embedded quotes, using the quotedValue option results in more bytes of input data
than the escapeChar option. When you have a choice, use unquoted syntax.
If you expect all values in all input fields (string or otherwise) to be uniformly enclosed in
quotes, then use the requireQuotes option to cause the system to enforce this usage. Using
the requireQuotes option improves the parsing overhead and provides extra robustness.
3-10 20525 Rev.3

Option Details
RecordDelim
Specifies that the row/record delimiter to be used is the string literal. Valid values must be
a maximum 8-byte UTF-8 string.
RecordLength
Specifies the length of the entire record. Includes the length itself, but does not include
the RecordDelimiter.
RemoteSource
Specifies the source datafile is remote, and takes the following values: ODBC, JDBC,
OLE-DB, or empty string. External tables created with the remote source set to ODBC,
JDBC, or OLE-DB are usable only through those, respectively. External tables created with
the remote source not set (or set to empty string) are usable from any client (the source
datafile path is assumed to be on the Netezza host, even if the load/unload is initiated
remotely from a different host).
Note that nzsql does not support remote loads/unloads using external tables (you can only
create external tables remotely), though it does support loads/unloads locally on the host.
This option is automatically set to ODBC if the hostname option is set to anything but local-
host or the reserved IP address (127.0.0.1).
RequireQuotes
Specifies if quotes are mandatory. The default is false. If set to true, the quoted value must
be set to YES, SINGLE, or DOUBLE. See “QuotedValue” on page 3-10.
SkipRows
Specifies the number of initial rows to skip before loading the data. The default is 0 (none).
After the system has a candidate binary record from an input row, it determines whether to
insert that record into the target table:
 If you did not specify this option, the system inserts every record.
 If you specified this option and the input row counter is less than or equal to the
skipRows count, the system discards the candidate binary record (skipped). Otherwise,
the system inserts the record.
Note: If you use the skipRows option, the system skips that number of rows, and then
begins the count for the maxErrors and/or maxRows options (if you have specified them).
Note that this cannot be used for 'header' row processing in a datafile, as even the skipped
rows are processed first, so the data in the header rows should be valid with respect to the
external table definition.
This option can be used for doing a dry-run to validate the datafile is correct, before loading
into a user table, by setting a maximum value.
20525 Rev.3 3-11

SocketBufSize
Specifies the chunk size at which to read the data from the source file, expressed in bytes.
Valid values range from 64KB to 800MB, with a default value of 8MB. Values outside this
range result in a system notice that the value will be reset to the appropriate minimum or
maximum level. This is used to fine-tune the performance of loads, depending on the
speed at which the source data is available for loads.
TimeDelim
Specifies the single-byte character that separates the time components. The default is ':'.
 If you specify the timeDelim option as an empty string, you must specify the hour, min-
utes, and optional seconds as two-digit numbers.
 If you specify the 12-hour format, you can precede the AM or PM token with a single
space. Note that the tokens, AM and PM are case-insensitive.
The system checks syntax and range errors. If an error occurs, the system discards the
record to the nzbad file and logs an error with the record number in nzlog file.
TimeRoundNanos
Rounds the time value to six fractional seconds digits. You can use the timeRoundNanos
option to specify allowing but rounding non-zero digits with smaller than microsecond
precision.
 If you do not use the timeRoundNanos option, a value is accepted, as long as it can be
stored without loss of precision.
 If you specify this option, the value is accepted, even when full precision of any frac-
tional seconds cannot be preserved. In this case, the value is rounded.
For example, consider the following timestamps:
1999/12/31 23:59:59.9999994
1999/12/31 23:59:59.9999995
Both of these timestamps specify finer than microsecond resolution. Without the
option, each would be rejected. Using the option, the first sample timestamp would
round to:
1999/12/31 23:59:59.999999
The second sample would round to:
2000/01/01 00:00:00.0
Note: This option is not supported for Fixed-Length format, and is also referred to as the
TimeExtraZeros option.
TimeStyle
Specifies the time format (‘24HOUR’, ‘12HOUR’) used in the data file. The default is
‘24HOUR’.
TruncString
Specifies truncating a string and inserting it into the declared storage.
False – Default, the system reports an error when a string exceeds its declared storage.
3-12 20525 Rev.3

Option Processing
True – Truncate any string value that exceeds its declared char/varchar storage.
Y2Base
If you specify the Y2-style date, use the -y2Base option to specify the start of the 100-year
range. Table 3-5 provides some examples of date ranges and their corresponding input
values.
Table 3-5: The -y2Base Option
Desired Range 1900…1999 1923…2022 1976…2075 2000…2999
Option -y2Base 1900 -y2Base 1923 -y2Base 1976 -y2Base 2000
In Y2 input
00 1900 2000 2000 2000
01 1901 2001 2001 2001
02 1902 2002 2002 2002
24 1924 1924 2024 2024
25 1925 1925 2025 2025
76 1976 1976 1976 2076
77 1977 1977 1977 2077
98 1998 1998 1998 2098
99 1999 1999 1999 2099
Option Processing
This section contains additional information on how the system processes the options.
Counting Rows
The system uses a line-oriented input format – one line of text is an input row. It operates
by isolating successive rows in the input stream. Every time it finds a new row, it incre-
ments a row counter (starting with number 1) and analyzes the contents of the row.
During analysis two sorts of errors can occur:
 The input text may not match the expected format.
 A field value might fail to meet a requirement imposed by the target table schema.
If a row contains no errors, the system converts the row into a candidate binary record.
20525 Rev.3 3-13

Handling Bad Rows

When the system encounters an error, it stops analyzing the row, appends the row to the
bad rows file, writes a supporting diagnostic message to the nzlog file describing the posi-
tion and nature of the error, and increments a rejected rows counter.
Delineating Input Rows

Input rows are separated by any of the common end-of-line conventions: <CR><LF>,
<LF><CR>, <CR>, or <LF>. In UNIX environments <LF> is commonly known as NewLine.
The last row/line need not have an end-of-line character.
Neither of the pairs <CR><CR> nor <LF><LF> is a valid end-of-line sequence. Instead
each pair encloses an empty row containing no values. The system considers such an
empty row valid only if you specified the fillRecord option, and you specified that every col-
umn in the target tables is capable of being set to null.
Matching Input Fields to Table Columns

The system determines the shape of input rows by inspecting the target table’s schema.
The fields are paired-up left-to-right with the columns in the target schema. Once the sys-
tem has located the start of a field, the declared type of the corresponding target column
guides further processing.
Note: It is an error for a row to contain more fields than the target table contains columns.
Using String and Non-string Fields

If an input field corresponds to a column declared char, nchar, varchar, or nvarchar, the
system considers it a string field, with all other types as non-string fields. This distinction is
important because spaces are significant within string fields, but not elsewhere.
Note: An empty field or a field containing only spaces can represent a legitimate string
value, but can never be a legitimate non-string value.
The system uses the following rules based on whether the field is a string field:
 If the field is a string field – All characters from the beginning of the field to the termi-
nating delimiter or end of row sequence contribute to the field’s value.
 If the field is a non-string field – The system skips any leading spaces, interprets or
converts the field’s contents, and skips any trailing spaces.
The string/non-string distinction also affects the details of how a field indicates that it is
null. For more information, see “Handling the Absence of a Value” on page 3-14.
Handling the Absence of a Value

In SQL, a record must include a value if a column is declared not null. When a record con-
tains no value for a column, the column is considered to be null. The system provides an
explicit and implicit method for conveying nullness.
 The explicit method includes a specific token in the field instead of a value. By
default, this token is the word “null” (case insensitive). You can use the nullValue
option to change this token to any other 1-4 character alphabetic token. You can pre-
cede or follow an occurrence of the explicit null token in a non-string field with
adjacent spaces. For the system to recognize an explicit null token in a string field, the
3-14 20525 Rev.3

Option Processing
token cannot have preceding or trailing adjacent spaces. The explicit null token
method makes it impossible to express a string consisting of exactly the text of the null
token.
 The implicit method interprets an empty field as null. This method is always available
to non-string fields independent of any nullValue option setting and works even if the
non-string field contains spaces. You can use the implicit method on string fields only
if you have set the nullValue option to the empty string ('').
The system considers a string field empty (potentially null) only if it contains truly zero
characters (no spaces). Setting nullValue to the empty string makes it impossible to set
any character varying (alias varchar(n)) column to an empty, zero-length string. In other
words, if the system encounters an empty string and the nullValue is set to '', then the
system treats the empty string as a null value.
Enabling Load Continuation

If you enable load continuation with the allowReplay option, or set the session variable
LOAD_REPLAY_REGION to true, the system ensures that a simple load using external
tables has the ability to continue after the system has been paused and resumed. You do
not have to abort and resubmit the load. If no value is specified for the allowReplay option,
or n is 0, the system defaults to the postgres default setting. If n is a valid non-zero num-
ber, it specifies the number of allowable query restarts.
The system accomplishes this automatic resumption by holding records to be sent to the
SPU in the replay region in host memory. After the system sends the data in this region to
the SPUs, it does a partial commit that forces all the unwritten data to the SPUs’ disks and
allows the system to re-use the reload region’s data buffers. If an SPU reboots or resets, the
system rollbacks to the last partial commit, and reprocesses and resends the data.
Note: Setting this option has a performance impact which depends on the speed of the
incoming data. In addition, system memory is used for the data buffering that enables
loads to be continued. When the buffer memory is exhausted, new loads will pend until
needed memory becomes available.
Load continuation cannot operate on any table that has one or more materialized views in
an active state. Before enabling load continuation, suspend the associated materialized
views. You can suspend active materialized views either through the NzAdmin tool or by
issuing the ALTER VIEWS command. Sample syntax for ALTER VIEWS follows.
ALTER VIEWS ON <table> MATERIALIZE SUSPEND
Once loading has completed, you can update and activate the materialized views for the
table. Sample syntax follows.
ALTER VIEWS ON <table> MATERIALIZE REFRESH
For more information, see the IBM Netezza System Administrator’s Guide.
Handling Legal Characters

Input is composed of the printing characters (bytes 33-255), space (byte 32), horizontal
tab (byte 9), line feed (byte 10) and carriage return (byte 13). By default you cannot use
the nonprinting control characters.
 Specify the ctrlChars option to permit control characters (bytes 1-8, 11-12, and
14-31) to appear within strings. In this case, only 0, 10, and 13 are not allowed.
20525 Rev.3 3-15

 Specify the crInString option to permit unescaped carriage returns (cr) in char/varchar
fields. If you specify the crlnString option, line feed (LF) becomes the default
end-of-row indicator.
 Specify the escapeChar option to permit any character preceded with a backslash (\) to
be interpreted as an escape character. In this way, you could use the zero (byte 0), line
feed (byte 10), carriage return (byte 13), or the closing delimiter.
 Specify the ignoreZero option to cause the system to check every character for zero.
This causes the system to skip over each zero it finds and to consider the next charac-
ter. If you specify this option, you cannot include a zero byte in a string.
For example, assume <nul> is a null byte, the field delimiter is '|' and you have speci-
fied ignoreZero:
..|<nul>AB<nul>CDEF<nul>|..
fills a char(6) column with 'ABCDEF'.
..|<nul>127<nul>|..
fills a byteint column with binary 01111111 (= 0x7F).
Table 3-6 lists the end-of-row and control characters that are permitted with the different
nzload system options. The  mark indicates that the option is specified or allowed.
Table 3-6: Control Characters and End of Record Characters
Control Characters Allowed within

Options End of Record
Strings
-crlnString -ctrlChars lf cr crlf lfcr 0 1-8 ht lf 11 12 cr 14-31
    
   
         
        
Note: In Fixed-Length format, control characters are treated differently. For more informa-
tion, see Chapter 6, “Using Fixed-Length Format.”
Session Variables
The following session variables work as nzload options.
 LOAD_REPLAY_REGION – See “Enabling Load Continuation” on page 3-15.
 MAX_QUERY_RESTARTS – The number of restarts allowed for load continuation. See
“Enabling Load Continuation” on page 3-15.
 LOAD_LOG_MAX_FILESIZE – The maximum allowed size in MB for the log file.
3-16 20525 Rev.3

CHAPTER 4
Using nzload
 How the nzload Command Works
 Using the nzload Command
 Configuration File Example
This chapter describes the nzload command. Netezza SQL is the Netezza Structured Query
Language (SQL), which runs on the Netezza data warehouse appliance. Throughout this
document, the term SQL refers to Netezza’s SQL implementation. For nzload usage exam-
ples, see Appendix A, “Examples and Grammar.”
How the nzload Command Works

The nzload command is a SQL CLI client application that allows you to load data from the
local host or a remote client, on all the supported client platforms. The nzload command
processes command-line load options to send queries to the host to create an external table
definition, run the insert/select query to load data, and when the load completes, drop the
external table.
The nzload command connects to a database with a user name and password, just like any
other Netezza appliance client application. The user name specifies an account with a par-
ticular set of privileges, and the system uses this account to verify access.
Note: While you can use the nzload command as an ODBC client application, it does not
require nor does it work with Data Source Name (DSN). It bypasses the ODBC Driver Man-
ager and connects directly to the Netezza ODBC driver.
Protection and Privileges

To run the nzload command, you must have the CREATE EXTERNAL TABLE privilege and
access privileges to that table or database (LIST, INSERT, SELECT). For more information,
see the IBM Netezza System Administrator’s Guide.
Note: If you issue the nzload command from the Netezza appliance host itself, and the user
issuing the command is not the user nz, you must do one of the following:
 Ensure that the GROUP nz has READ permissions for the data file to load.
 Use the -host option with the nzload command (such as nzload -host <hostname>).
4-1
Concurrency and Transactions

You can run multiple nzload jobs in parallel with each job adding records to the same
tables. While loading, you can run concurrent queries, inserts, updates, and deletes against
committed records in the target tables.
The nzload command performs all insertions into the target table within a single transac-
tion. The nzload command commits the transaction at the end of the job, provided it does
not detect any fatal errors. Only after the nzload command has committed the transaction
are the newly loaded records visible to other queries. When encountering a load error while
running multiple concurrent loads, only the load with the error does not complete.
While the nzload job is running, it sends records to the SPUs along with the current trans-
action ID. When an SPU receives new records, it immediately allocates resources and
writes the records to the database or the table on the disk.
If the nzload command cannot commit the transaction, these storage resources remain
allocated. To free up this disk space, use the nzreclaim command on the specific table or
database. For more information about the nzreclaim command, see the IBM Netezza Sys-
tem Administrator’s Guide.
If you cancel an nzload job, the nzload command does not commit the transaction.
Program Invocation
The nzload command is a command-line program that accepts input values from multiple
sources. The precedence order is the following:
 Command line
 Control file. Without a control file, you can only do one load at a time, and using a con-
trol file allows multiple loads. See “Using a Control File” on page 4-5.
 Environmental variables (only used for user, password, database, and host)
 Built-in defaults
Option names are case insensitive. Every option has a standard name for use in either the
command line or the control file. For more information about the input values, see
Table 4-1 on page 4-3.
Many options include a token argument, which you can enclose in either single or double
quotes. The nzload command treats alphabetic characters in option token arguments as
case-insensitive (for example -boolStyle YES_NO is equivalent to -boolStyle yes_no).
Note: You must quote options that require a punctuation character as a token, and use an
escape character if quotes are part of the argument.
Using the nzload Command

The nzload command takes options and arguments. You can accept the defaults or specify
options on the command line, in the control file, or through environment variables. For a
complete listing of all options, see Appendix C, “Option Names.”
Syntax
The nzload command uses the following syntax:
nzload [-h|-rev] [options]
4-2 20525 Rev.3

Inputs
The nzload command uses many of the options for external tables, as detailed in
Chapter 3, “External Table Options.” Particular options for nzload are shown in Table 4-1.
Table 4-1: The nzload Options
Option Description
-cf filename Specifies the control file. For more information, see “Using a Con-
trol File” on page 4-5.
-df filename Specifies the datafile to load. If you do not specify a path, the sys-
tem uses the special token <stdin> to store the filepath string.
Corresponds to the DataObject external table option.
-lf filename Specifies the log file name. If the file exists, this appends to it.
-bf filename Specifies the bad/rejected rows filename (overwrite if the file
exists).
-outputDir dir Specifies the output directory for the log and bad/rejected rows
files. Corresponds to the LogDir external table option.
-logFileSize n Session variable (LOAD_LOG_MAX_FILESIZE) that specifies the

size (in MB) of the log and bad/rejected rows files. The default is
2000MB (2GB).
-fileBufSize Specifies the chunk size (MB for fileBufSize or bytes for fileBufBy-
-fileBufByteSize teSize) at which to read the data from the source file. Corresponds
to the SocketBufSize external table option.
-allowReplay Session variables (LOAD_REPLAY_REGION and MAX_QUERY_

-allowReplay n RESTARTS) that specify the number of query restarts for load con-
tinuation if a SPU has reset or failed over. If n is a valid non-zero
number, it specifies the number of allowable query restarts. If no
value is specified, or n is 0, the system defaults to the postgres
default setting.
Additional Options
The nzload takes the following additional options:
Table 4-2: nzload Additional Options
Option Description
-u user Specifies the logon user name [NZ_USER].
-pw password Specifies the user’s password [NZ_PASSWORD].
-host name Specifies the hostname or IP address [NZ_HOST]. Runs on the

local host if not specified here. If you set this to any name but
localhost or any IP address but the reserved one (127.0.0.1), the
system sets the remotesource option to ODBC.
20525 Rev.3 4-3

Table 4-2: nzload Additional Options
Option Description
-caCertFile path Specifies the pathname of the root CA certificate file on the cli-
ent system. This argument is used by Netezza clients who use
peer authentication to verify the Netezza host system. The default
value is NULL which skips the peer authentication process.
-securityLevel level Specifies the security level that you want to use for the session.
The argument has four values:
• 0 – preferredUnsecured – This is the default value. Specify
this option when you would prefer an unsecured connection,
but you will accept a secured connection if the Netezza sys-
tem requires one.
• 1 – onlyUnsecured – Specify this option when you want an
unsecured connection to the Netezza system. If the Netezza
system requires a secured connection, the connection will be
rejected.
• 2 – preferredSecured – Specify this option when you want a
secured connection to the Netezza system, but you will accept
an unsecured connection if the Netezza system is configured
to use only unsecured connections.
• 3 – onlySecured – Specify this option when you want a
secured connection to the Netezza system. If the Netezza sys-
tem accepts only unsecured connections, or if you are
attempting to connect to an Netezza system that is running a
release prior to 4.5, the connection will be rejected.
Note: If you specify an invalid value for the -securityLevel argu-
ment of the nzload command, the command defaults to the
preferredUnsecured (0) level.
-db database Specifies the database to load [NZ_DATABASE].
-t table Specifies the table name. You can specify a fully qualified name
for this value.
-port Specifies the port to use, allowing you to override the default.
-loginTimeout Allows you to enter a different value, expressed in seconds, for

<int-seconds> the login timeout. This overrides the default value of 30 seconds.
Outputs
The nzload command exits with the following codes:
 0 – Successful, all input records were inserted.
 1 – Failed, no records were inserted due to an error or errors found during the load.
 2 – Successful, but errors found during the input did not exceed the error threshold
(-maxErrors), good records were inserted.
4-4 20525 Rev.3

Using a Control File

An nzload control file allows you to define load operations in a text file without having to
specify the options on the nzload command line. You can also use control files to run mul-
tiple concurrent loads, with different options into one command instance. Each load is a
different transaction, and in case of a rollback, any completed transactions are good.
Options
Within a control file, you can specify the following options:
 Any of the valid options for an external table. For more information, see Appendix C,
“Option Names.” You can specify the long format name of the option or the short for-
mat name.
 Database – Specifies the name of the database to load.
 Table – Specifies the name of the table to load the data.
 Badfile (bf) – Specifies the name of the nzbad file, which contains any records which
could not be loaded. The default is table.database.nzbad.
 Logfile (lf) – Specifies the name of the nzload log file, which contains messages and
errors that occurred during the load processing. The default is table.database.nzlog.
 Datafile – Specifies the pathname of the file that you want to load into the specified
table and database. The datafile option must be the first line of the control file, fol-
lowed by list of control file options in curly braces {}. You can specify more than one
datafile, each with its own set of options, in the control file.
 Decimal delimiter – Specifies to use a comma instead of a period as a decimal delim-
iter. The default delimiter is a period.
The options in a control file are case-insensitive. For example, you could specify the option
in letter formats such as database, DataBase, Database, or DATABASE.
Note that command line options take precedence over any equivalent options specified in a
control file. This allows you to override any control file options as necessary without chang-
ing the control file. If you specify a control file for the nzload command, you cannot specify
a data file argument (-df) on the command line.
Syntax
The syntax for using a control file is as follows, where each sequence can be another load:
DATAFILE <filename>
{
[<option name> <option value>]*
}
For example, the following control file options load the data from customer.dat into the cus-
tomer table:
DATAFILE /home/operation/data/customer.dat
{
Database dev
TableName customer
}
20525 Rev.3 4-5

If you save the control file contents as a text file (named cust_control.txt in this example)
you can specify it using the nzload command as follows:
nzload -cf /home/nz/sample/cust_control.txt
Load session of table 'CUSTOMER' completed successfully
When you use the nzload command, note that you cannot specify both the -cf and -df
options in the same command. You can load from a specified data file, or load from a con-
trol file, but not both in one command.
The following control file options define two data sets to load. Note that the options can
vary for each data set.
DATAFILE /home/operation/data/customer.dat
{
Database dev
TableName customer
Delimiter '|'
Logfile operation.log
Badfile customer.bad
}
DATAFILE /home/imports/data/inventory.dat
{
Database dev
TableName inventory
Delimiter '#'
Logfile importload.log
Badfile inventory.bad
}
If you save these control file contents as a text file (named import_def.txt in this example)
you can specify it using the nzload command as follows:
nzload -cf /home/nz/sample/import_def.txt
Load session of table 'CUSTOMER' completed successfully
Load session of table 'INVENTORY' completed successfully
Configuration File Example

The following is an example of a fixed format configuration file.
{
outputdir /home/nzuser
crinstring 'true'
ctrlchars 'true'
decimaldelim '.'
format fixed
recordlength 10
maxerrors 0
tablename refnull
layout ( fld1 bool 1_0 bytes 1 , fld2 char(5) bytes 5 , fld3 char(4)
bytes 4)
}
4-6 20525 Rev.3

CHAPTER 5
Unloading Data
 Unloading Options
 Unloading Data to a Remote Client System
This chapter describes the options for unloading data. For usage examples, see
Appendix A, “Examples and Grammar.”
Unloading Options
The following external table options are not supported for unloads. For a complete list of
external table options, see Chapter 3, “External Table Options.”
 CtrlChars
 FillRecord
 IgnoreZero
 Layout
 LogDir
 MaxErrors
 MaxRows
 QuotedValue
 RecordDelim
 RecordLength
 RequireQuotes
 SkipRows
 TimeRound Nanos/TimeExtraZeros
 TruncString
 Y2Base
The IncludeZeroSeconds external table option is used only for unloads. The 2-digit format
of the DateStyle external table option is not supported for unloads.
5-1
Unloading Data to a Remote Client System

A special use of the CREATE EXTERNAL TABLE/INSERT INTO commands is to stream data
from an Netezza database file on an Netezza host system to a remote client. This unload
does not remove rows from the database, but rather stores the unloaded data in a flat file
that is suitable for loading back into an Netezza database.
You can unload data to any of the supported Netezza appliance clients, which include Win-
dows, Linux, Solaris, AIX, and HP-UX (PA-RISC and Itanium). You can unload all data types
(including Unicode) and file types (uncompressed and compressed formats).
Note: You must be the admin user or have the Create External Table administration privi-
lege to create an external table, and you must have permission to write to the data object’s
path. Both the Netezza appliance host and remote client must be running Netezza release
3.1 or greater.
Note: Unloading for Fixed-Length format is not supported.
To unload to a remote client, do the following:

1. Establish an ODBC or JDBC connection between the client machine and the Netezza
appliance host. For example on a Linux or UNIX client, type:
>isql
2. Use the CREATE EXTERNAL TABLE command to create an external table. An example
follows:
CREATE EXTERNAL TABLE emp_backup SAMEAS emp USING (
DATAOBJECT ('/tmp/emp.dat')
REMOTESOURCE 'ODBC');
INSERT INTO emp_backup SELECT * FROM emp;
In the example, note that the DATAOBJECT file specification must be a valid file on the
receiving machine. REMOTESOURCE must be either ODBC, JDBC, or OLE-DB. The
ODBC/JDBC/OLE-DB client must be connected with the corresponding Netezza appli-
ance library. If you do not specify a remote source, the system unloads the data to a file
on the Netezza appliance host.
3. To reload the data in the external table, you can use a SQL query such as:
INSERT INTO emp SELECT * FROM emp_backup;
Verify that emp is empty before you reload the data.
5-2 20525 Rev.3

CHAPTER 6
Using Fixed-Length Format
 Formatting Background
 Format Options
 Layout Definitions
 Building the Fixed-Length Format Definition
This chapter describes the fixed-length format for loading data into external tables.
Formatting Background
All data is a series of byte-sequences and has an associated data type, used here as a con-
ceptual or abstract attribute of the data. Without an associated data type, a byte-sequence
can be interpreted in too many ways.
A single data type can be represented in different forms. For example, an integer data type
can be represented or stored in various types of binary format, or in human-readable
text/character format (typically ASCII). Similarly, dates, times and other data types have
multiple representations used by different programs, languages, and environments. At
some point, though, these data types must be represented in readable form, so users can
do something with the data. Data for loading into the data warehouse typically is presented
in either delimited format or fixed-length format, using either ASCII or UTF-8.
Fixed-Length Format
Fixed-length format files use ordinal positions, which are offsets to identify where fields are
within the record. There are no field delimiters, and there may be no end-of-record delim-
iter. Data in fixed-length format files seldom has decimal or time delimiters, as these are
not necessary, and take up space. Because the fields are fixed in size, the location of
delimiters are fixed, and can are specified in the layout definition, which accompanies the
fixed-length format data file.
Loading fixed format data into the database requires that you define the target data type for
the field, as well as the location within the record.
Not all fields in a fixed-length format file need to be loaded, and can be skipped using the
‘filler’ specification. The order of fields in the data file must match the order of the target
table, or an external table definition must be defined, which specifies the order of the
fields as database columns. Using an external table definition in combination with an
insert-select statement allows field order to be changed.
6-1
Unknown or null values are typically represented by known data patterns, which are classi-
fied as representing null. The Netezza system identifies and act on these values.
Data Attributes
The typical data attributes in fixed-length format files are as follows:
 Data Type – The data at a given offset in a record is always of the same type.
 Representation – The representation is constant, and each field has a fixed width. Data
within a field is always presented in the same way. Certain items such as radix points,
time separators, and date delimiters are always at the same place and are typically
implied, rather than being actually present in the data file.
 Value – The value can be an actual value or a null indicator. Data representations
which indicate a null value are specified by the layout definition. Assuming null is
allowed.
 Length – There is no length specification within the data file, as length in the file is
fixed, and the length attribute is specified by the layout definition.
 Null-ness – Null-ness is identified in the layout definition as either a specific data pat-
tern, such as “all spaces” or as being “flagged” by a value in another column.
Format Options
For the fixed-length format, new options have been added, and some have been changed.
New Options
The following added external table options are valid only for the fixed-length format.
 RecordLength – The length of the entire record, including null-indicator bytes (if any)
and excluding record-delimiter (if any).
 No default value
 Constant integer
 RecordDelim – The row/record delimiter.
 Default is ‘\n’ (new-line). Note that the field is literally interpreted, so ‘\n’ looks for
those characters, and not ‘new-line’
 The end-of-record delimiter is entered between single quotes. The end-of-record
indicator can be up to a maximum 8 bytes long
 The omission of a record delimiter is defined by side-by side single quotes
 Layout – Mandatory for fixed-length format. Used to define the location of fields of the
input record.
 No default value
 Comma separated zone definitions within braces
6-2 20525 Rev.3

Format Options
Changed Options
The following external table options have a different meaning for the fixed-length format:
Table 6-1: Changed Option Meanings
Option Meaning
CtrlChars Text-Delimited: If False (default), unescaped control characters (except \t)

error out.
Exception: If CtrlChars is False and CrInString is True, \r (carriage Return)
can be used without error.
If True, unescaped control characters \0 and \n error out (also \r if CrIn-
String is False).
Fixed-Length: If True, all unescaped characters allowed.
If False (default), unescaped characters error out.
Exceptions: \t, \n (and \r if CrInString is ON).
CrInString Text-Delimited: Augments CtrlChars behaviors.

Fixed-Length: Used only when CtrlChars is OFF.
MaxErrors Sets the maximum number of allowed (non-fatal) errors before aborting
the load. Since the parser now reports errors for each field or zone rather
than just one error for the row, multiple errors can be reported for the
same row, so this limit must be set accordingly. When the parser sees an
error in a field/zone, it recovers (using the field/zone length) and contin-
ues from the next field/zone, until the End-of-Record, a fatal error, or this
maxerrors limit is reached.
Fatal errors include the following:
• RecordLength mis-match
• RecordDelimiter not found
• RecordLength invalid (negative values or zero)
• Zone length invalid (negative values)
• UTF-8 initial byte is invalid
• UTF-8 continuation bytes are invalid
Unsupported Options
The following external table options are not supported for fixed-length format, and if set,
result in an error:
 Encoding
 FillRecord
 IgnoreZero
 TimeExtraZeros
 TruncString
 AdjustDistZeroInt
 IncludeZeroSeconds
20525 Rev.3 6-3

 Delimiter
 EscapeChar
 QuotedValue
 RequireQuotes
Default Values
The following existing external table options work as default values for zone definitions:
 NullValue – Default for the ‘NULLIF’ clause of all zones.
 DateStyle, DateDelim, TimeStyle, TimeDelim, BoolStyle – Default for zone style for cor-
responding date, time and bool zones.
Layout Definitions
Layout is an ordered collection of zone (field) definitions, and is a required option for
fixed-length format. Each zone (field) definition is made up of mutually exclusive
(non-overlapping) clauses. These clauses must be in the following order, although some are
optional and can be empty:
 Use-type – Indicates whether a zone is a normal (data) zone or a filler zone. For data
zones, this value is omitted. Filler zones can only be specified in bytes. Other use-types
exist, but are not used for fixed-length format data.
 Name – The name of the zone. Duplicate zone names are not allowed. This definition is
not currently used, but is typically provided to identify the field.
 Type – Defines the zone type. When not specified, type is defaulted to the correspond-
ing table column’s type. Filler-zones must have a zone type of INT. Valid values are as
follows:
 CHAR
 VARCHAR
 NCHAR
 NVARCHAR
 INT1
 INT2
 INT4
 INT8
 INT
 UINT1
 UINT2
 UINT4
 UINT8
 UINT
6-4 20525 Rev.3

Layout Definitions
 FLOATING
 DOUBLE
 NUMERIC
 BOOL
 DATE
 TIME
 TIMESTAMP
 TIMETZ
 Style – Defines the zone representation, and is optional. This is defaulted based on the
zone-type and ‘Format’ option. All other styles are only valid for their corresponding
non-textual zone-types. Valid values are the following:
 INTERNAL – Valid only for textual zones (CHAR/VARCHAR/NCHAR/NVARCHAR)
 DECIMAL – Valid for integer/numeric zone types
 DECIMALDELIM – Valid for numeric, float, double, and time-styles (time, timetz,
and timestamp) zone type
 FLOATING – Valid for float or double zone type
 SCIENTIFIC – Valid for float or double zone type
 YMD <‘date-delim’> (and other date-styles currently supported in external table
options DateStyle and DateDelim; valid for date zones
 12Hour <’time-delim’> (and other time-styles currently supported in external table
options TimeStyle and TimeDelim; valid for time zones)
 24Hour <’time-delim’> (and other time-styles currently supported in external table
options TimeStyle and TimeDelim; valid for time zones)
 YMD <’date-delim’> 24Hour <’time-delim’> (and other combinations of date and
time styles currently supported for external table options DateStyle, DateDelim,
TimeStyle and TimeDelim; valid for timestamp and timetz zones
 TRUE_FALSE, Y_N, 1_0 (and other boolean styles currently supported for external
table option BoolStyle; valid for boolean zones). Style has to be in accordance with
format
 Length – Specified in bytes.
 Nullif – Defines the zone null-ness attribute. For fixed format files, this clause speci-
fies a known data pattern within the field which when present signifies the field is null.
Length should be equal to or less than the column width, and maximum length is 39
bytes.
Nulls are detailed in Table 6-2:
Table 6-2: Layout Example
Use Type Name Type Style Length Unholiness
NA f1 Int4 DECIMAL Bytes 10 Nullify @ = 0
20525 Rev.3 6-5

Table 6-2: Layout Example
Use Type Name Type Style Length Unholiness
NA f2 Date YMD Bytes 10 Nullify &=

‘2000-10-10’
NA f3 Char(20) INTERNAL Chars 10 Nullify && =”
Filler f4 Char(4) NA Bytes 10 NA
Building the Fixed-Length Format Definition

Fixed-length format files must have a format definition. This section shows examples of
typical fixed-length format definitions for typical data types.
End-of-Record
When fixed format records end in a on-line character, no action is required, On-line is the
default end-of-record delimiter. When there is no record separator, use single quotes side
by side, as in the following example:
RecordDelim ‘’
RecordDelim is a literal sequence of up to 8 bytes, which does not translate common
escape representations or support functions like CHAR(8).
Record Length
Record Length is optional, but can provide feedback that the format definition has the cor-
rect length. This excludes the end-of-record delimiter. The following is an example:
Recordlength NNN
Skipping Fields
The following clause skips four bytes:
“filler char(4) bytes 4”
However, the preferred method is to indicate the field being skipped, as in the following
example:
“filler fld_name char(4) bytes 4”
Temporal Values
Temporal values in fixed-length format files often omit delimiters. Table 6-3 shows clauses
that load dates, times, and timestamps without delimiters.
Table 6-3: Temporal Values
Datatype Value Format Clause
Date 20101231 date1 date YMD'' bytes 8
Time 231559 time1 time(6) 24hour '' bytes 6
6-6 20525 Rev.3

Building the Fixed-Length Format Definition
Table 6-3: Temporal Values
Timestamp 0101231231559 stamp1 timestamp(6) 24hour '' bytes 14
Timestamp 2010123123155 (Load as char(24), then use insert-select)

9000001 to_timestamp(col,’YYYYMMDDHH24MISSUS')
Date 2010-12-31 date2 date YMD'-' bytes 10
Time 23.15.59 time2 time(6) 24hour '.' bytes 8
Timestamp 2010-12-31 tms2 timestamp(6) YMD ‘-’ 24hour ':' bytes 19

23:15:59
Timestamp 2010-12-31 tms3 timestamp(6) YMD ‘-’ 24hour ':' bytes 26

23:15:59.0001
Timetz 12:30:45+03:00 Tz1 TIMETZ(6) 24HOUR ':' bytes 14
Timetz 123045+-0300 (Load as char(11) then use insert-select)

(substring(col1,1,2)||':'|| substring(col1,3,2)||':'||sub-
string(col1,5,5)||':'|| substring(col1,10,2))::timetz
Numeric Values
Table 6-4 shows numeric values.
Table 6-4: Numeric Values
Integer 32767 int1 int2 bytes 5
Int8 9123456789123456 int2 int8 bytes 16
Numeric 2315.59 num1 numeric(6,2) bytes 7
Numeric 231559 (Load as char(6) then use insert-select)

(col/100)::numeric(6,2)
Floating 1.2345678 flt1 floating bytes 9
Floating 12345678 (Load as char(8) then use insert-select)

(substring(col1,1,1)||'.'||sub-
string(col1,2,7))::float
Double 1.2345678 flt1 double bytes 9
Double 12345678 (Load as char(8) then use insert-select)

(substring(col1,1,1)||'.'||sub-
string(col1,2,7))::double
20525 Rev.3 6-7

Logical Values
Table 6-5 shows logical values.
Table 6-5: Logical Values
Boolean Y or y, N or n BOOL Y_N BYTES 1
Boolean 1, 0 BOOL 1_0 BYTES 1
Boolean T or t, F or f BOOL T_F BYTES 1
Null Values
Fixed-length format files typically use ‘magic’ values to represent nulls. Adding a nullif
clause to any specification allows the column to be checked for null. A nullif clause has the
following parts:
 The keyword “nullif”
 The column reference
 The test expression
As an example, a file specification where field1 is a date and is considered null if it has the
value ’99991231’ would have the following characteristics:
 The nullif specification would be as follows:
“nullif &=’99991231’”
 The entire specification would be as follows:
“fld1 date YMD'' bytes 8 nullif &=’99991231’”
 All format specifications support the nullif clause.
In addition to &=, which evaluates to ‘string must exactly match,’ the nullif clause also
supports &&=, which allows substring matching. This is useful in cases where the string
may occur anywhere in a field with space padding. For example nullif &&=’N’ matches the
different expressions “ N “, “N “, “ N”.
Table 6-6 shows null values:
Table 6-6: Null Values
Datatype Null Value Format Clause
Boolean ‘ ‘ (1 space) BOOL Y_N BYTES 1 NULLIF &=’ ‘ (1 space)
DATE 000000 DATE YMD ‘’ BYTES 6 NULLIF &=’000000’
INT ‘ ‘ (6 spaces) INT BYTES 6 NULLIF &=’ ‘ (6 spaces)
6-8 20525 Rev.3

APPENDIX A
Examples and Grammar
What’s in this appendix
 The nzload Command
 Reference Examples
 Decimal Delimiter Examples
 SQL Grammar
 Fixed-Length Format Definition
 Script Example For Loading Data Using Fixed-Length Format
This appendix includes examples for using external tables, the nzload command, SQL
grammar, and references.
The nzload Command

The following examples describe how to specify nzload arguments, how to use named
pipes, and sample ways of using nzload.
Specifying nzload Arguments

The following examples show how to specify the admin and password or accept the
defaults.
 To load the table repeat_cust, delimited by |, and specifying the input file click-
stream.dat, enter:
nzload -t repeat_cust -delim '|' -df clickstream.dat
This example uses default values for the -u, -pw, -db option. For more information
about these default values, see Table 4-1 on page 4-3.
 To load the database dev as user admin with the password production, specifying the
table name areacode, using tab delimiters, and specifying the input file as phone-pre-
fix.dat, enter:
nzload -u admin -pw production -db dev -t areacode -delim '\t' -df
phone-prefix.dat
Note: To ensure optimum performance, run the GENERATE STATISTICS command after
you have loaded a table or database. For more information about the generate statistics
command, see the IBM Netezza System Administrator’s Guide.
A-1
Using Named Pipes

To load a large amount of data, use a named pipe to stream the data to external tables or to
the nzload command. The nzload command loads the data as it fills the pipe, and does not
exit until it receives the end-of-file indicator. Note that stdin is supported for nzload.
To use a named pipe to load tables with the nzload command, do the following:
1. Create a zero-length, named pipe file, using the Linux command mkfifo:
mkfifo mypipe
2. Do the following in a background session:
nzload -db <my_db> -t my_table -delim “|” -df /export/home/<my_db>/
mypipe
3. Do the following in a foreground session:
cat /export/home/nz/<my_db>/my_table.dat > mypipe
Sample nzload Usage

The following provides some sample nzload usage.
 To specify the name of the load file, enter:
nzload -u admin -pw password -host nzhost -db emp -t name -df /tmp/
daily/Import.bad
 To specify the boolean style, enter:
nzload -u admin -pw password -host nzhost -db emp -t name -df /tmp
-boolStyle yes_no
 To specify the name of the control file, enter:
nzload -u admin -pw password -host nzhost -db emp -t name -cf /tmp/
daily/control.file
 To allow unescaped carriage returns in char() and varchar() fields, enter:
-crinString
 To allow an ASCII value 1-31 in char() and varchar() fields, enter:
nzload -u admin -pw password -host nzhost -db emp -t name
-ctrlChars
 To specify the delimiter to use with the dateStyle option, enter:
-dateDelim ‘/’ -dateStyle MDY
 To specify how to interpret the date format, enter:
-dateDelim ‘/’ -dateStyle MDY’
 To specify the field delimiter, enter:
nzload -u admin -pw password -host nzhost -db emp -t name -delim
‘,’
 To specify using an escape character, enter:
A-2 20525 Rev.3

The nzload Command
-escapeChar ’\\’
 To specify an input line with fewer columns than the table definition, enter:
-fillRecord
 To specify discarding the byte value zero in the char() and varchar() fields, enter:
-ignoreZero no
 To specify the log file name, enter:
nzload -u admin -pw password -host nzhost -db emp -t name -lf /tmp/
daily/import.log
 To specify the maximum number of errors, enter:
-maxErrors 100
 To specify stopping processing when the specified number of records are in the data-
base, enter:
nzload -u admin -pw password -host nzhost -db emp -t name -maxRows
100
 To specify the string to use for the null value, enter:
-nullValue ‘none’
 To specify the output directory for the log files, enter:
-outputDir /tmp/daily
 To specify that quotes are mandatory, except for null values, enter:
-requireQuotes quoted value YES
 To specify the delimiter to use for time formats, enter:
-timeDelim ‘.’
 To specify allowing but rounding non-zero digits with smaller than microsecond resolu-
tion, enter:
-timeRoundNanos
 To specify the time style value in the data file, enter:
-timeStyle 12hour
 To specify truncation a string and inserting it into the declared string, enter:
-truncString
 To specify the first year in the YY format, enter:
nzload -u admin -pw password -host nzhost -y2Base 2000
20525 Rev.3 A-3

 To enable load continuation, enter:

-allowReplay
Reference Examples
Examples for references are as follows:
Table A-1: Reference Examples
Reference Meaning
BYTES &2 Error only internal @ reference allowed for length-clause (in any
format/zone-type).
BYTES @ An error length-clause cannot refer itself.
NULLIF & = ‘123’ Self-reference (no number) is valid in null-clause.

The length has to be BYTES/CHARS 3, for text-styles.
Matches (nullif evaluates to ‘true’) ONLY ‘123’ (a row in the
external file containing ‘123’).
NULLIF && = ‘123’ Matches (nullif evaluates to ‘true’) ‘123’, ‘ 123 ‘ ‘ 123 ‘, if
SPACE is skipped.
Length has to be at least BYTES 3 (text-styles) or BYTES 4.
NULLIF @ = 123 Valid for numerical zones.

Matches ‘123’, ‘ 123 ‘ and so on, in text format, with spaces
skipped.
NULLIF @ = ‘2000- Valid for date zones

01-01’
Decimal Delimiter Examples

The following are examples of how to use the new decimal delimiter option. The option use
is shown in bold text.
For text-delimited format for the table level:
INSERT INTO <target-table> SELECT * FROM ‘<external-table>’ USING
(delim ‘delim’ decimalDelim ‘,’);
For fixed-length format for the table level:
(decimalDelim ‘,’ format ‘fixed’ layout (c1 int bytes 4, c2 float bytes
6, c3 numeric(10,2) bytes 11, c4 time 24HOUR ‘:’ bytes 11 );
For fixed-length format for the column level.
 For numeric data type:
(format ‘fixed’ layout (c1 int bytes 4 , c2 float bytes 6, c3
numeric(10,2) decimal ‘,’ bytes 11) );
A-4 20525 Rev.3

SQL Grammar

(format ‘fixed’ layout (c1 int bytes 4 , c2 float bytes 6, c3
numeric(10,2) decimal decimalDelim ‘,’ bytes 11) );
 For float data type:

(format ‘fixed’ layout (c1 int bytes 4 , c2 float floating ‘,’
bytes 6, c3 numeric(10,2) bytes 11) );
(format ‘fixed’ layout (c1 int bytes 4 , c2 float floating
decimalDelim ‘,’ bytes 6, c3 numeric(10,2) bytes 11) );
 For double data type:
(format ‘fixed’ layout (c1 int bytes 4 , c2 double exponential ‘,’
(format ‘fixed’ layout (c1 int bytes 4 , c2 float exponential
 For time data types (time, timetz, timestamp):
(format ‘fixed’ layout (c1 int bytes 4 , c2 time 12HOUR
(format ‘fixed’ layout (c1 int bytes 4 , c2 time timeDelim ‘-’
(format ‘fixed’ layout (c1 int bytes 4 , c2 time timeDelim ‘-’ ‘,’
(format ‘fixed’ layout (c1 int bytes 4 , c2 time 12HOUR ‘-’
(format ‘fixed’ layout (c1 int bytes 4 , c2 time 12HOUR ‘-’ ‘,’
(format ‘fixed’ layout (c1 int bytes 4 , c2 time 12HOUR timeDelim
‘-’ decimalDelim ‘,’ bytes 12, c3 numeric(10,2) bytes 11) );
(format ‘fixed’ layout (c1 int bytes 4 , c2 time 12HOUR timeDelim
‘-’ ‘,’ bytes 12, c3 numeric(10,2) bytes 11) )
SQL Grammar
This section provides an explanation of the SQL grammar used for CREATE EXTERNAL
TABLE.
[INSERT INTO <normal-table>] SELECT <col-list> FROM EXTERNAL [name] ‘<data-file>’
[USING ‘(‘ <Load-options>’)’]
20525 Rev.3 A-5

CREATE EXTERNAL TABLE <ext-table-name><External-table-shape> (<External-table-

shape> | SAMEAS <tablename>) USING ‘(‘ <Load-options> ‘)’
CREATE EXTERNAL TABLE [name] ‘file path’ [USING ‘C’ load-options ‘)’ AS SELECT-
statement
Load-options: Load-option
| Load-option Load-options // space separated list of USING clause options
Load-option: FORMAT TEXT | INTERNAL | FIXED
| RECORDLENGTH <n>| Length-ref-expr
| RECORDDELIM <string-literal-max-8-bytes >
| LAYOUT ( Zone-definitions )
…..
Zone-definitions: Zone-def
| Zone-def ‘,’ Zone-definitions // comma-separated lists of zone definitions
Zone-def: [Zone-use-type] [Zone-name] [Zone-type] [Zone-style] [Zone-len] [Nullness]
Zone-use-type: REF | FILLER
Zone-name: Identifier
Zone-type: CHAR| VARCHAR
| NCHAR| NVARCHAR
| BOOL
| INT1 | INT2 | INT4 | INT8 | INT
| UINT1 | UINT2 | UINT4 | UINT8 | UINT
| NUMERIC
| FLOATING| DOUBLE
| DATE | TIME | TIMESTAMP | TIMETZ
Zone-style: INTERNAL
| DECIMAL [‘decimal-delim’]
| FLOATING | SCIENTIFC [‘decimal-delim’]
| Date-format
| Time-format
| Date-format Time-format
Date-format:
| DateStyle [‘date-delim’]
| DATE DELIM ‘date-delim’
Time-format:
A-6 20525 Rev.3

Fixed-Length Format Definition
| TimeStyle [‘time-delim’] [‘decimal-delim’]

| TIME DELIM ‘time-delim’ DecimalDelim ‘decimal-delim’
Date-style: YMD| DMY | MDY |.. // all date styles

Time-style: 12HOUR | 24HOUR
Zone-len: BYTES <n> | <Length-ref-expr>

| CHARACTERS <n> | <Length-ref-expr>
Zone-ref: External-ref
| Isolated-ref
| Internal-ref
External-ref: &[n] // 1 based absolute position of zones, 0, negative values for relative posi-
tions backwards
Isolated-ref: &&[n] // 1 based absolute position of zones 0, negative values for relative posi-
tions backwards
Internal-ref: @[n]// 1 based absolute position of zones, 0, negative values for relative posi-
tions backwards
Length-ref-expr: Internal-ref [ Operator <n> ]
Operator: + | -
Fixed-Length Format Definition

The following is a sample data record:
20011228YF2001122814313425 Forest St Marlborough MA017525083828200600
The record is defined by the following column layout:
 Columns 1-8 Date format YYYYMMDD Null when value is ‘99991231’
 Column 9 Boolean Y/N Null when value is space ‘ ‘
 Column 10 Boolean T/F Null when value is space ‘ ‘
 Column 11-24 Time stamp format YYYYMMDDHHMMSS Null when value is
‘99991231000000’
 Column 25-39 Character Address Null when value is all spaces
 Column 40-52 Character City Null when value is ’****NULL*****‘
 Column 53-54 Character State Null when value is ‘##’
 Column 55-59 Number zipcode Null when value is all zeroes
 Column 60-68 Character Phone Null when value is all zeroes
 Column 69-72 Number(3,2) Example 600 would be 6.00 Never Null
 Column 73 Newline end of record.
20525 Rev.3 A-7

The following is an example of the Netezza External Table definition for this data:
CREATE EXTERNAL TABLE sample_ext (
Col01 DATE ,
Col09 BOOL ,
/* Skipped col10 */
Col11 TIMESTAMP,
Col26 Char(12),
Col38 Char(10),
Col48 Char(2),
Col50 Int4,
Col56 CHAR(10),
Col67 CHAR(3) /* Numeric(3,2) cannot be loaded directly */
)
USING (
dataobject('/home/test/sample.fixed')
logdir '/home/test'
recordlength 72 /* does not include end of record delimiter */

recorddelim '
' /* This is actually a newline between the single quotes, really not needed as newline is
default */
format 'fixed'
layout (
Col01 DATE YMD '' bytes 8 nullif &='99991231',
Col09 BOOL Y_N bytes 1 nullif &=' ',
FILLER Char(1) Bytes 1, /* was col10 space */
Col11 TIMESTAMP YMD '' 24HOUR '' bytes 14 nullif &='99991231000000',
Col26 CHAR(15) bytes 15 nullif &=' ', /* 15 spaces */
Col38 CHAR(13) bytes 13 nullif &='****NULL*****' ,
Col48 CHAR(2) bytes 2 nullif &='##' ,
Col50 INT4 bytes 5 nullif &='00000' ,
Col56 CHAR(10) bytes 10 nullif &='0000000000',
Col67 CHAR(3) bytes 3 /* We cannot load this directly, so we use an insert-select */
) /* end layout */
); /* end external table definition. */
A-8 20525 Rev.3

Script Example For Loading Data Using Fixed-Length Format
INSERT INTO sampleTable

SELECT
Col01,
Col09,
Col11,
Col26,
Col38,
Col48,
Col50,
Col56 ,
(Col67/100)::numeric(3,2) as Col67 /* convert char to numeric(3,2) */
FROM sample_ext ;
Script Example For Loading Data Using Fixed-Length Format

The following is an example of a script to load data using the Fixed-length format.
LOGDIR="/tmp"
DIR="/tmp"
NZSQL="nzsql -db test -c"
function CreateDb()
{
nzsql -c "create database test"
}
function CleanUp()
{
$NZSQL "drop table textDelim_tbl"
$NZSQL "drop table textFixed_tbl"
}
function CreateTable()
{
$NZSQL "create table textDelim_tbl(col1 int, col2 char(10), col3
date)"
$NZSQL "create table textFixed_tbl(col1 int, col2 char(10), col3
date)"
}
function CreateDataFile()
{
20525 Rev.3 A-9

# Create text delimited data file

cat > $DIR/delimData << EOF
1|Customer|12/7/2011
2|Netezza|02/16/2010
EOF
# Create text fixed data file

cat > $DIR/fixedData << EOF
1HelloWorld2011-12-07
2Netezza 2010-02-16
EOF
}
function LoadData()
{
# nzload using text format
nzload -t textDelim_tbl -df $DIR/delimData -db test -outputDir
$LOGDIR -delim '|' -dateStyle MDY -dateDelim '/'
#nzload using fixed format
nzload -t textFixed_tbl -df $DIR/fixedData -db test -outputDir
$LOGDIR -format fixed -layout "col1 int bytes 1, col2 char(10) bytes
10, col3 date YMD '-' bytes 10"
function UnloadData()
{
$NZSQL "insert into textDelim_tbl select * from external '$DIR/
delimData' using (Delimiter '|' DateStyle 'MDY' DateDelim '/');"
}
CreateDb
CleanUp
CreateTable
CreateDataFile
LoadData
UnloadData
A-10 20525 Rev.3

APPENDIX B
Troubleshooting
 Tips for Successful Loading
 nzload Error Handling
This section contains examples to aid you in troubleshooting data loading.
Tips for Successful Loading

The following sections describe how to analyze your data, how to set up loading and how to
troubleshoot any problems that might arise.
Create Your Table

Before you create your table, check the following:
 Choose a distribution key. If you know the primary key or a column that is used fre-
quently in joins, use that one. Use a distribution key with the highest selectivity. For
more information about distribution keys, see the IBM Netezza System Administrator’s
Guide.
 Check that any column that does not contain null data (or should not contain null data)
is declared as not null. The system processes not null columns more quickly.
 Check if you have number fields. Are they declared as int8, int4, smallint, byteint or
numeric(s,p). The smaller the storage, the better for large tables.
Determine Your Data Format

Consider the following when determining the format of your data:
 Check how many data fields there are in each input line of the data file. Are there the
same number of columns defined in the target-table definition?
 If there are fewer columns than fields, is it acceptable to extend the schema to
have filler columns? If not, then the load will not succeed.
 If there are more columns than fields, is it acceptable to use null values to insert
into those columns? If it is acceptable, specify the -fillRecord option.
B-1
 Check the field delimiter. It should be a character used to separate one field value from
another. This field delimiter should be unique and should not appear in a field value,
especially in a char or varchar string. Use the -delim option to specify the field
delimiter.
 Check whether there are any NULL values in the data source. How is the null value
expressed in the data file? The RDBMS industry convention is to use the string “null”
to represent a null value. If the data file uses a different representation, use the -
nullValue option to override the default null value. The new value can be an empty
string or a value in the range of a-z or A-Z and no longer than four characters.
 Check whether there are any date, time, time with time zone, or timestamp data types
in the table schema. If there are, what style is the date value? The style of these data
type values must be consistent throughout the nzload job.
 Check the handling of string fields for char() or varchar() data types. Does the longest
or largest value fit into the storage of the char() or varchar() declaration? If not, is it
possible to alter the schema to accommodate the longest string?
 If schema cannot be altered, is truncating a string an acceptable solution?
 If truncation is acceptable, specify the -truncString option.
 If neither is acceptable, the nzload command treats the record with the long string
as an error record. The nzload command discards the record to the nzbad file and
logs an error with the record and column numbers in nzlog file.
 See whether there are any special characters used in the string fields. For example,
CR, CRLF, or a character in a string that is the same as the field delimiter? This
violates the unique character rule.
 If there are special characters, can you regenerate the data file to have an escape
character added to these special characters? If so, then use the -escapeChar '\\'
option to process the strings.
 If you cannot regenerate the data file, then the load will contain incomplete and
invalid records.
Consider the Load Source

See whether you are using pipes. If so, are they from another local feed or from across a
network? The preferred method is to read from a named pipe, rather than to read from
stdin/stdout.
Look at the file. Is the file on an NFS mounted directory? If so, remember that your load
performance is constrained by the speed of the network.
Run the Job

Make a copy of your source table before beginning the load if you are running on a produc-
tion system. Making a backup is extremely fast within the Netezza appliance and is better
than reloading from a backup. For example, the syntax for making a copy is as follows:
CREATE TABLE loan_backup AS SELECT * FROM loan;
Stage the data before moving it to a production system. Create a new table, load it, validate
it, then use the ALTER TABLE command to move the tables to production. For example:
B-2 20525 Rev.3

Tips for Successful Loading
ALTER TABLE loan RENAME TO loan_lastmonth;

ALTER TABLE loan_stage RENAME TO loan;
If you are running multiple nzload jobs to load into a table, use unique names for your
nzbad files. The nzload command generates the default file name by using the <table-
name>.<databasename> and appending the extension .nzbad. Loading into the data table
of the dev database uses the default file name "data.dev.nzbad" for the nzbad file. Each
instance of the nzload command overwrites the existing file. If you want to preserve the bad
records that are stored in this file, use the -bf <file name> option to specify a different
name for each nzload job.
Note: If your default system case is uppercase, the system displays lowercase table names
as uppercase in nzlog files, for example, DATA.DEV.nzlog and DATA.DEV.nzbad
Run the Linux top command on the host to monitor CPU resources. Consider running more
loads concurrently if resources are available.
Troubleshoot
If you see the error message, “Too many data fields for table,” use the Linux command
head -1 on the data file to get the first row, which may contain the column’s names
extracted. Compare these to your create table's DDL and see if their physical positions
match.
If you see the error message, “Data type mismatch on column 5,” use the Linux command
cut -d^ -f 5 inputfile | more to look at the individual data values in the source file and then
compare them to your DDL. Compare these to your create table's DDL and see if their phys-
ical positions match.
Handle Exceptions
Repeat the load on the -bf file. If there are many exceptions, fix them and re-extract from
the source system. If they are few, use a text editor to change data. To make large substitu-
tions, use the Linux sed or awk commands.
Validate the Results

After the load completes, validate the results by comparing them with the source system.
Count the number of rows and select min/max/sum of each numeric and min/max of each
date column in the table.
Generate Statistics
Remember to run the generate statistics command on your tables and/or database after you
have loaded new data.
Test Performance
If your data is evenly distributed, you should see peak loading performance of at least 75
percent CPU utilization on the host. You can monitor utilization by running the Linux top
command during the load. If you see less CPU utilization that means either the data is
skewed so that all SPUs are not sharing the workload or the parser is waiting for data.
If your input data is skewed, that is, all records are being sent to a small number of SPUs,
those SPUs become the performance bottleneck.
20525 Rev.3 B-3

If your CPU utilization is less than 75 percent and the data is evenly distributed, you might
have a streaming problem:
 If the load is running from the local host, determine the source of the data.
Look for other concurrent database activities — such as activities that are SPU-to-SPU
broadcast intensive or SPU disk I/O intensive.
 If the data is not locally staged or is on a SAN / NFS mount, determine if the bottle-
neck is the remote source of the data or the network.
The performance of the Netezza appliance system depends on the number of SPUs. If,
however, data is being streamed across an external network, then the performance is
limited by the speed of the network.
Test the network by using the FTP command to send a file between the source and the
local host, and measure the transfer rate. Under optimal conditions, a Gig-E network
transfers at a rate of ~1000Mb/second, or ~125MB/second or ~450GB/hour.
nzload Error Handling

The nzload command does extensive error checking. This section describes how the nzload
command interprets different data types and the way it handles syntax errors.
Reporting Errors
The nzload command returns standard error status when it completes.
 0 – The load was successful, all input records were inserted.
 1 – The load failed, no records were inserted due to error(s) found during load.
 2 – The load was successful, but the system found error in input that did not exceed
error threshold (-maxErrors), so good records were inserted.
The nzload command writes high-level errors to the terminal (stderr), nzlog file, and nzbad
file. You can specify the nzlog and nzbad filenames on the command line or through the
use of a control file. For more information, see “Using a Control File” on page 4-5.
Note: Periodically delete log files to free up disk space.
Understanding nzload Log Files

The system creates the following nzlog file as the result of the command line:
nzload -u admin -pw password -t member_profile -db dev -maxErrors 10 -delim '\t'
 -maxErrors allows the nzload command to continue processing until it has found 10
errors.
 -delim '\t' specifies the TAB delimiter.
The system appends to the nzlog file for every nzload command that loads the same table
into the same database. The system names the nzlog file based on the table and the data-
base name with the extension .nzlog. So, in this example, the file name is:
member_profile.dev.nzlog
There is also a member_profile.dev.nzbad file that contains the record(s) that caused the
error(s). The system overwrites this file each time you invoke the nzload command for the
same table and database name (unlike the behavior of the nzlog file).
B-4 20525 Rev.3

APPENDIX C
Option Names
 Specifying Options
This section details the different methods of using options.
Specifying Options
Table C-1 shows how to enter the external table options when using the command line
method (used for nzload), in a control file, or as part of a SQL command.
Table C-1: Specifying External Table Options
Option Command Line Control File SQL
AllowReplay -allowreplay NA LOAD_REPLAY_

REGION
MAX_QUERY_
RESTARTS
BadFile -bf badfile NA
BoolStyle -boolStyle boolstyle BOOLSTYLE
Compress -compress compress COMPRESS
CRinString -crInString crinstring CRINSTRING
CtrlChars -CtrlChars ctrlchars CTRLCHARS
Database -db database NA
Datafile -df datafile DATAOBJECT
DateDelim -dateDelim datedelim DATEDELIM
DateStyle -dateStyle datestyle DATESTYLE
DecimalDelim -decimaldelim decimaldelim DECIMALDELIM
Delimiter -delim delim DELIM

-delimiter delimiter DELIMITER
C-1
Encoding -encoding encoding ENCODING
EscapeChar -escape escape ESCAPE

-escapeChar escapechar ESCAPECHAR
FillRecord -fillRecord fillrecord FILLRECORD
Format -format format FORMAT
IgnoreZero -ignoreZero ignorezero IGNOREZERO
IncludeZeroSec- NA NA INCLUDEZEROSEC-
onds ONDS
Layout -layout layout LAYOUT
LogDir -outputDir outputdir LOGDIR
LogFile -lf logfile NA
LogFileSize -logFileSize NA LOAD_LOG_MAX_

FILESIZE
MaxErrors -maxErrors maxerrors MAXERRORS
MaxRows -maxRows maxrows MAXROWS
NullValue -nullValue nullvalue NULLVALUE
QuotedValue -quotedValue quotedvalue QUOTEDVALUE
RecordDelim -recdelim recdelim RECDELIM
RecordLength -reclength recordlength RECLENGTH
RemoteSource -host NA REMOTESOURCE
RequireQuotes -requireQuotes requirequotes REQUIREQUOTES
SkipRows -skipRows skiprows SKIPROWS
SocketBufSize -fileBufSize socketbufsize SOCKETBUFSIZE

-fileBufByteSize
SuspendMviews -suspendMviews NA NA
Tablename -t tablename NA
TimeDelim -timeDelim timedelim TIMEDELIM
C-2 20525 Rev.3

Specifying Options
TimeRound -timeRoundNanos timeroundnanos TIMEROUNDNANOS

Nanos -timeExtraZeros timeextrazeros TIMEEXTRAZEROS
TimeExtraZeros
TimeStyle -timeStyle timestyle TIMESTYLE
TruncString -truncString truncstring TRUNCSTRING
Y2Base -y2Base y2base Y2BASE
20525 Rev.3 C-3

C-4 20525 Rev.3

APPENDIX D
Notices and Trademarks
 Notices
 Trademarks
 Electronic Emission Notices
 Regulatory and Compliance
This section describes some important notices, trademarks, and compliance information.
Notices
This information was developed for products and services offered in the U.S.A.
IBM may not offer the products, services, or features discussed in this document in other
countries. Consult your local IBM representative for information on the products and ser-
vices currently available in your area. Any reference to an IBM product, program, or service
is not intended to state or imply that only that IBM product, program, or service may be
used. Any functionally equivalent product, program, or service that does not infringe any
IBM intellectual property right may be used instead. However, it is the user's responsibility
to evaluate and verify the operation of any non-IBM product, program, or service.
IBM may have patents or pending patent applications covering subject matter described in
this document. The furnishing of this document does not grant you any license to these
patents. You can send license inquiries, in writing, to: This information was developed for
products and services offered in the U.S.A.
IBM Director of Licensing
IBM Corporation
North Castle Drive
Armonk, NY 10504-1785 U.S.A.
For license inquiries regarding double-byte (DBCS) information, contact the IBM Intellec-
tual Property Department in your country or send inquiries, in writing, to:
IBM World Trade Asia Corporation
Licensing 2-31 Roppongi 3-chome, Minato-ku
Tokyo 106-0032, Japan
The following paragraph does not apply to the United Kingdom or any other country where
such provisions are inconsistent with local law: INTERNATIONAL BUSINESS MACHINES
CORPORATION PROVIDES THIS PUBLICATION "AS IS" WITHOUT WARRANTY OF ANY
KIND, EITHER EXPRESS OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE
D-1
IMPLIED WARRANTIES OF NON-INFRINGEMENT, MERCHANTABILITY OR FITNESS FOR

A PARTICULAR PURPOSE. Some states do not allow disclaimer of express or implied war-
ranties in certain transactions, therefore, this statement may not apply to you.
This information could include technical inaccuracies or typographical errors. Changes are
periodically made to the information herein; these changes will be incorporated in new edi-
tions of the publication. IBM may make improvements and/or changes in the product(s)
and/or the program(s) described in this publication at any time without notice.
Any references in this information to non-IBM Web sites are provided for convenience only
and do not in any manner serve as an endorsement of those Web sites. The materials at
those Web sites are not part of the materials for this IBM product and use of those Web
sites is at your own risk.
IBM may use or distribute any of the information you supply in any way it believes appropri-
ate without incurring any obligation to you.
Licensees of this program who wish to have information about it for the purpose of
enabling: (i) the exchange of information between independently created programs and
other programs (including this one) and (ii) the mutual use of the information which has
been exchanged, should contact:
IBM Corporation
Software Interoperability Coordinator, Department 49XA
3605 Highway 52 N
Rochester, MN 55901
U.S.A.
Such information may be available, subject to appropriate terms and conditions, including
in some cases, payment of a fee.
The licensed program described in this document and all licensed material available for it
are provided by IBM under terms of the IBM Customer Agreement, IBM International Pro-
gram License Agreement or any equivalent agreement between us.
Any performance data contained herein was determined in a controlled environment.
Therefore, the results obtained in other operating environments may vary significantly.
Some measurements may have been made on development-level systems and there is no
guarantee that these measurements will be the same on generally available systems. Fur-
thermore, some measurements may have been estimated through extrapolation. Actual
results may vary. Users of this document should verify the applicable data for their specific
environment.
Information concerning non-IBM products was obtained from the suppliers of those prod-
ucts, their published announcements or other publicly available sources. IBM has not
tested those products and cannot confirm the accuracy of performance, compatibility or
any other claims related to non-IBM products. Questions on the capabilities of non-IBM
products should be addressed to the suppliers of those products.
All statements regarding IBM's future direction or intent are subject to change or with-
drawal without notice, and represent goals and objectives only.
All IBM prices shown are IBM's suggested retail prices, are current and are subject to
change without notice. Dealer prices may vary.
This information contains examples of data and reports used in daily business operations.
To illustrate them as completely as possible, the examples include the names of individu-
als, companies, brands, and products. All of these names are fictitious and any similarity to
the names and addresses used by an actual business enterprise is entirely coincidental.
D-2 20525 Rev.3

Trademarks
COPYRIGHT LICENSE:
This information contains sample application programs in source language, which illustrate
programming techniques on various operating platforms. You may copy, modify, and distrib-
ute these sample programs in any form without payment to IBM, for the purposes of
developing, using, marketing or distributing application programs conforming to the appli-
cation programming interface for the operating platform for which the sample programs are
written. These examples have not been thoroughly tested under all conditions. IBM, there-
fore, cannot guarantee or imply reliability, serviceability, or function of these programs.
Each copy or any portion of these sample programs or any derivative work, must include a
copyright notice as follows:
© your company name) (year). Portions of this code are derived from IBM Corp. Sample
Programs.
© Copyright IBM Corp. _enter the year or years_.
If you are viewing this information softcopy, the photographs and color illustrations may not
appear.
Trademarks
IBM, the IBM logo, ibm.com and Netezza are trademarks or registered trademarks of Inter-
national Business Machines Corporation in the United States, other countries, or both. If
these and other IBM trademarked terms are marked on their first occurrence in this infor-
mation with a trademark symbol (® or ™), these symbols indicate U.S. registered or
common law trademarks owned by IBM at the time this information was published. Such
trademarks may also be registered or common law trademarks in other countries. A current
list of IBM trademarks is available on the Web at “Copyright and trademark information” at
ibm.com/legal/copytrade.shtml.
Adobe is a registered trademark of Adobe Systems Incorporated in the United States, and/
or other countries.
Linux is a registered trademark of Linus Torvalds in the United States, other countries, or
both.
Microsoft, Windows, Windows NT, and the Windows logo are trademarks of Microsoft Corpo-
ration in the United States, other countries, or both.
NEC is a registered trademark of NEC Corporation.
UNIX is a registered trademark of The Open Group in the United States and other
countries.
Java and all Java-based trademarks are trademarks of Sun Microsystems, Inc. in the United
States, other countries, or both.
Red Hat is a trademark or registered trademark of Red Hat, Inc. in the United States and/or
other countries.
D-CC, D-C++, Diab+, FastJ, pSOS+, SingleStep, Tornado, VxWorks, Wind River, and the
Wind River logo are trademarks, registered trademarks, or service marks of Wind River Sys-
tems, Inc. Tornado patent pending.
APC and the APC logo are trademarks or registered trademarks of American Power Conver-
sion Corporation.
Other company, product or service names may be trademarks or service marks of others.
20525 Rev.3 D-3

Electronic Emission Notices

When you attach a monitor to the equipment, you must use the designated monitor cable
and any interference suppression devices that are supplied with the monitor.
Federal Communications Commission (FCC) Statement

Note: This equipment has been tested and found to comply with the limits for a Class A
digital device, pursuant to Part 15 of the FCC Rules. These limits are designed to provide
reasonable protection against harmful interference when the equipment is operated in a
commercial environment. This equipment generates, uses, and can radiate radio frequency
energy and, if not installed and used in accordance with the instruction manual, may cause
harmful interference to radio communications. Operation of this equipment in a residential
area is likely to cause harmful interference, in which case the user will be required to cor-
rect the interference at his own expense.
Properly shielded and grounded cables and connectors must be used in order to meet FCC
emission limits. IBM is not responsible for any radio or television interference caused by
using other than recommended cables and connectors or by unauthorized changes or mod-
ifications to this equipment. Unauthorized changes or modifications could void the user's
authority to operate the equipment.
This device complies with Part 15 of the FCC Rules. Operation is subject to the following
two conditions: (1) this device may not cause harmful interference, and (2) this device
must accept any interference received, including interference that might cause undesired
operation.
Industry Canada Class A Emission Compliance Statement

This Class A digital apparatus complies with Canadian ICES-003.
Avis de conformité à la réglementation d'Industrie Canada

Cet appareil numérique de la classe A est conforme à la norme NMB-003 du Canada.
Australia and New Zealand Class A Statement

Attention: This is a Class A product. In a domestic environment this product may cause
radio interference in which case the user may be required to take adequate measures.
European Union EMC Directive Conformance Statement

This product is in conformity with the protection requirements of EU Council Directive
2004/108/EC on the approximation of the laws of the Member States relating to electro-
magnetic compatibility. IBM cannot accept responsibility for any failure to satisfy the
protection requirements resulting from a nonrecommended modification of the product,
including the fitting of non-IBM option cards.
Attention: This is an EN 55022 Class A product. In a domestic environment this product
may cause radio interference in which case the user may be required to take adequate
measures.
Responsible manufacturer:
International Business Machines Corp.
New Orchard Road
Armonk, New York 10504
914-499-1900
D-4 20525 Rev.3

Electronic Emission Notices
European Community contact:

IBM Technical Regulations, Department M456
IBM-Allee 1, 71137 Ehningen, Germany
Telephone: +49 7032 15-2937
Email: tjahn@de.ibm.com
Germany Class A Statement

Deutschsprachiger EU Hinweis: Hinweis für Geräte der Klasse A EU-Richtlinie zur Elektro-
magnetischen Verträglichkeit
Dieses Produkt entspricht den Schutzanforderungen der EU-Richtlinie 2004/108/EG zur
Angleichung der Rechtsvorschriften über die elektromagnetische Verträglichkeit in den EU-
Mitgliedsstaaten und hält die Grenzwerte der EN 55022 Klasse A ein.
Um dieses sicherzustellen, sind die Geräte wie in den Handbüchern beschrieben zu instal-
lieren und zu betreiben. Des Weiteren dürfen auch nur von der IBM empfohlene Kabel
angeschlossen werden. IBM übernimmt keine Verantwortung für die Einhaltung der Schut-
zanforderungen, wenn das Produkt ohne Zustimmung der IBM verändert bzw. wenn
Erweiterungskomponenten von Fremdherstellern ohne Empfehlung der IBM gesteckt/einge-
baut werden.
EN 55022 Klasse A Geräte müssen mit folgendem Warnhinweis versehen werden:
“Warnung: Dieses ist eine Einrichtung der Klasse A. Diese Einrichtung kann im Wohnbere-
ich Funk-Störungen verursachen; in diesem Fall kann vom Betreiber verlangt werden,
angemessene Maßnahmen zu ergreifen und dafür aufzukommen.”
Deutschland: Einhaltung des Gesetzes über die elektromagnetische Verträglichkeit von Geräten
Dieses Produkt entspricht dem “Gesetz über die elektromagnetische Verträglichkeit von
Geräten (EMVG)”. Dies ist die Umsetzung der EU-Richtlinie 2004/108/EG in der Bundes-
republik Deutschland.
Zulassungsbescheinigung laut dem Deutschen Gesetz über die elektromagnetische Verträglichkeit von Geräten
(EMVG) (bzw. der EMC EG Richtlinie 2004/108/EG) für Geräte der Klasse A
Dieses Gerät ist berechtigt, in Übereinstimmung mit dem Deutschen EMVG das EG-Konfor-
mitätszeichen - CE - zu führen.
Verantwortlich für die Einhaltung der EMV Vorschriften ist der Hersteller:
International Business Machines Corp.
New Orchard Road
Armonk, New York 10504
914-499-1900
Der verantwortliche Ansprechpartner des Herstellers in der EU ist:
IBM Deutschland
Technical Regulations, Department M456
IBM-Allee 1, 71137 Ehningen, Germany
Telephone: +49 7032 15-2937
Email: tjahn@de.ibm.com
Generelle Informationen:
Das Gerät erfüllt die Schutzanforderungen nach EN 55024 und EN 55022 Klasse A.
20525 Rev.3 D-5

Japan VCCI Class A Statement
This is a Class A product based on the standard of the Voluntary Control Council for Inter-
ference (VCCI). If this equipment is used in a domestic environment, radio interference
may occur, in which case the user may be required to take corrective actions.
Japan Electronics and Information Technology Industries Association (JEITA) Statement
Japan Electronics and Information Technology Industries Association (JEITA) Confirmed

Harmonics Guidelines (products less than or equal to 20 A per phase)
Japan Electronics and Information Technology Industries Association (JEITA) Statement
Japan Electronics and Information Technology Industries Association (JEITA) Confirmed

Harmonics Guidelines (products greater than 20 A per phase)
Korea Communications Commission (KCC) Statement
This is electromagnetic wave compatibility equipment for business (Type A). Sellers and
users need to pay attention to it. This is for any areas other than home.
Russia Electromagnetic Interference (EMI) Class A Statement
People's Republic of China Class A Electronic Emission Statement
D-6 20525 Rev.3

Regulatory and Compliance
Taiwan Class A Compliance Statement
Regulatory and Compliance

Regulatory Notices
Install the NPS system in a restricted-access location. Ensure that only those trained to
operate or service the equipment have physical access to it. Install each AC power outlet
near the NPS rack that plugs into it, and keep it freely accessible.
Provide approved circuit breakers on all power sources.
Product may be powered by redundant power sources. Disconnect ALL power sources
before servicing.
High leakage current. Earth connection essential before connecting supply. Courant de
fuite élevé. Raccordement à la terre indispensable avant le raccordement au réseau.
Homologation Statement
This product may not be certified in your country for connection by any means whatsoever
to interfaces of public telecommunications networks.
Further certification may be required by law prior to making any such connection. Contact
an IBM representative or reseller for any questions.
WEEE
Netezza Corporation is committed to meeting the requirements of the European Union (EU)
Waste Electrical and Electronic Equipment (WEEE) Directive. This Directive requires pro-
ducers of electrical and electronic equipment to finance the takeback, for reuse or
recycling, of their products placed on the EU market after August 13, 2005.
20525 Rev.3 D-7

D-8 20525 Rev.3

Index
Index
A E
allowreplay 4-3, C-1 encoding 3-7, C-2
attributes errors
data 6-2 nzload handling B-4
escape C-2
escapechar 3-7, C-2
B external table
about 2-1
backup
backup and restore 2-4
external tables 2-4
displaying information 2-2
nzload B-2
examples 2-15
badfile 4-5, C-1
options 3-1
best practices
parsing 2-3
external tables 2-13
privileges 2-1
bigint, integer type 2-6
restrictions 2-13
boolstyle 3-3, C-1
byteint, integer type 2-6
F
C fileBufByteSize 4-3
filebufbytesize C-2
character strings
fileBufSize 4-3
char 2-10
filebufsize C-2
varchar 2-10
fillrecord 3-8, C-2
column constraint 2-10
fixed point 2-7
compress 3-4, C-1
floating point 2-8
compressed binary 1-2
format 3-8, C-2
concurrency 4-2
format options 6-2
control file
formatting, background 6-1
using 4-5
counting rows 3-13
CREATE EXTERNAL TABLE
dropping an external table 2-15 H
examples 2-15 host 4-3, C-2
crinstring 3-4, C-1
ctrlchars 3-4, C-1
I
ignorezero 3-8, C-2
D includezeroseconds 3-8, C-2
data attributes 6-2 integer, type 2-6
data loading
components 1-1
formats 1-2 L
data types
layout 3-9
fixed-point 2-7
definitions 6-4
floating-point 2-8
legal characters 3-15
integer 2-7
load continuation 3-15
supported 2-6
load. See also nzload
temporal 2-11
LOAD_LOG_MAX_FILESIZE 4-3
database C-1
LOAD_REPLAY_REGION 4-3, C-1
datafile 4-5, C-1
loading, success tips B-1
dataobject 3-4
log files 2-2
datedelim 3-5, C-1
logdir 3-9, C-2
datestyle 3-5, C-1
logfile 4-5, C-2
decimaldelim 1-2, 3-2, 3-6, 6-5
size C-2
delim C-1
logfilesize 4-3
delimiter 3-6, C-1
Index-1
Index
M bad 3-14
counting 3-13
matching input fields 3-14 input 3-14
MAX_QUERY_RESTARTS 4-3, C-1 skipping 3-11
maxerrors 3-9, C-2
maxrows 3-9, C-2
S
N session variables 3-16
skip rows 3-11
NOT NULL 3-10 skiprows C-2
nullvalue 3-9, C-2 smallint, integer type 2-6
numerics 2-6 socketbufsize 3-12, C-2
nzload command SQL grammar A-5
backup B-2 string versus non-string 3-14
boolStyle 4-2 supported data types 2-6
error reporting B-4 suspendmviews C-2
examples A-1
inputs 4-3
privileges 4-1
program invocation 4-2
T
tablename C-2
specifyng arguments A-1
syntax 4-2 temporal data types 2-11
textfixed, using 6-1
tips B-1
timedelim 3-12, C-2
uncommitted jobs 4-2
timeextrazeros C-3
using 4-1
nzmigrate 1-1 timeroundnanos 3-12, C-3
timestamp 2-12
nzreclaim command
timestyle 3-12, C-3
nzload jobs 4-2
timetz 2-12
transactions, nzload jobs 4-2
O troubleshooting B-1
truncstring 3-12
options
changed 6-3
external table 3-1
names C-1
U
new 6-2 unloading
examples 2-16
processing 3-3
options 5-1
unsupported 6-3
outputdir 4-3, C-2 remote client 5-2
P V
pipes A-2 value absence 3-14
privileges, load session 4-1
Y
Q y2base 3-13, C-3
quotedvalue 3-10, C-2
Z
R zone definition, default values 6-4
zones
recdelim C-2
default values 6-4
recorddelim 3-11
recordlength 3-11
references
examples A-4
remote client, unloading 5-2
remotesource 3-11, C-2
requirequotes 3-11, C-2
rows
Index-2

Netezza Data Loading Guide PDF

Uploaded by

Copyright:

Available Formats

Netezza Data Loading Guide PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Netezza Data Loading Guide PDF

Uploaded by

Copyright:

Available Formats

What are external tables and what are they used for?

What are external tables and what are they used for?

What data types are supported for external tables?

What data types are supported for external tables?

IBM Netezza 6.0.

IBM Netezza Data Loading Guide

© Copyright IBM Corporation 2011, 2012.

3 External Table Options

6 Using Fixed-Length Format

Appendix A: Examples and Grammar

Appendix C: Option Names

Appendix D: Notices and Trademarks

Audience for This Guide

Purpose of This Guide

Topics See the following

Introduction to Data Loading Concepts and Terms Chapter 1, “Overview”

How to use External Tables Chapter 2, “External Tables”

Details on the nzload command Chapter 4, “Using nzload”

Details on unloading data Chapter 5, “Unloading Data”

Details on the Fixed-Length format Chapter 6, “Using Fixed-Length

Examples of commands, format, and usage Appendix A, “Examples and

Command and Task Tips Appendix B, “Troubleshooting”

Legal notices and trademark information Appendix D, “Notices and

Symbols and Conventions

Comments on the Documentation

Data Loading Components

Data Loading Formats

New Decimal Delimiter Option

1-2 20525 Rev.3

About External Tables

Displaying External Table Information

2-2 20525 Rev.3

20525 Rev.3 2-3

Backing Up and Restoring

Transient External Tables

CREATE EXTERNAL TABLE 'filename' [USING (external_table_options)]

2-4 20525 Rev.3

SELECT <column_list | *> FROM EXTERNAL 'filename' (schema_definition)

Explicit Schema Definition

Implicit Schema Definition

Exporting Data Using Transient External Tables

Remote Transient External Tables

20525 Rev.3 2-5

Supported Data Types

Table 2-1: Supported Data Types

Data Type Example Description

byteint 120 See “Integer Data Types” on page 2-7.

numeric -99.56 See “Fixed-Point Data Types” on

real –81293.35 See “Floating-Point Data Types” on

char (n) salary See “Character Strings” on page 2-10

varchar (n) I am John See “Character Strings” on page 2-10

boolean true An ASCII string that contains any of the

date 2002-02-04 The date is an exact four-byte data type.

time 01:59:45 See “Time” on page 2-11.

timestamp 2002-02-04 01:15:33 See “Timestamp” on page 2-12.

2-6 20525 Rev.3

Integer Data Types

Table 2-2: Integer Description

Description • Optional leading sign