20000802: IMGCOPY's CYCLE Keyword

Pepo Juega wrote:

> Hello,
>
> I'm having difficulties in handling image DataSets to hold selected
> hours. In my example I try to keep 'synoptic slots' for meteosat
> images at 00Z, 03Z, 06Z, 09Z, 12Z, 15Z, 18Z & 21Z, and discard the
> rest of the incoming slots.
>
> The first thing I did was to create a new DataSet holding eight posi-
> tions to keep one day of information, and then scheduled a IMGCOPY
> command to feed this DataSet from the 1/2 hour source like this:

--->(This is done after UnPNGing off LDM wich arrives every 1/2 hour)
---> Remember I only want to keep 'synoptic slots' 

> IMGCOPY SRC/METIR DEST/SYNSLOTS CYCLE=03:00
>
> I was expecting to get the following for IMGLIST DEST/SYNSLOTS.ALL
>  Pos Satellite/         Date       Time      Center   Band(s)
>      sensor                                 Lat  Lon
>  --- -------------  ------------  --------  ---- ---- ------------
>    1  METEOSAT7     31 JUL 00213  00:00:00    40    4 8
>    2  METEOSAT7     31 JUL 00213  03:00:00    40    4 8
>    3  METEOSAT7     31 JUL 00213  06:00:00    40    4 8
>    4  METEOSAT7     31 JUL 00213  09:00:00    40    4 8
>    5  METEOSAT7     31 JUL 00213  12:00:00    40    4 8
>    6  METEOSAT7     31 JUL 00213  15:00:00    40    4 8
>    7  METEOSAT7     31 JUL 00213  18:00:00    40    4 8
>    8  METEOSAT7     31 JUL 00213  21:00:00    40    4 8
>
> ...but I got:
>  Pos Satellite/         Date       Time      Center   Band(s)
>      sensor                                 Lat  Lon
>  --- -------------  ------------  --------  ---- ---- ------------
>    1  METEOSAT7     31 JUL 00213  01:00:00    40    4 8
>    2  METEOSAT7     31 JUL 00213  04:00:00    40    4 8
>    3  METEOSAT7     31 JUL 00213  07:00:00    40    4 8
>    4  METEOSAT7     31 JUL 00213  10:00:00    40    4 8
>    5  METEOSAT7     31 JUL 00213  13:00:00    40    4 8
>    6  METEOSAT7     31 JUL 00213  16:00:00    40    4 8
>    7  METEOSAT7     31 JUL 00213  19:00:00    40    4 8
>    8  METEOSAT7     31 JUL 00213  22:00:00    40    4 8
>
> This apparently offends the specified formula for positioning at
> the destination DataSet. Could you please advice?

MUG Help Desk wrote:
> 
> Hi Pepo-
> 
> I'm having some difficulty reproducing your problem.  I set up the
> scheduler to start just after 0Z and continue every 3 hours with the
> following command:
> 
> SKE 2000214 00:05:00 999999 3:00:00 "IMGCOPY MET/FD SYNOPTIC/IMAGES
> CYCLE=03:00 BAND=8
> 
> So far this is what I've gotten:
> 
> IMGLIST SYNOPTIC/IMAGES.ALL
> Image file directory listing for:SYNOPTIC/IMAGES
>  Pos Satellite/         Date       Time      Center   Band(s)
>      sensor                                 Lat  Lon
>  --- -------------  ------------  --------  ---- ---- ------------
>    1  METEOSAT7      1 AUG 00214  00:00:00            8
>    2  METEOSAT7      1 AUG 00214  03:00:00            8
>    3  METEOSAT7      1 AUG 00214  06:00:00            8
>    4  METEOSAT7      1 AUG 00214  09:00:00            8
>    5  METEOSAT7      1 AUG 00214  12:00:00            8
>    6  METEOSAT7      1 AUG 00214  15:00:00            8
> IMGLIST: done
> 
> Possibly your commands were not run at the correct time and the CYCLE=
> got out of sequence somehow?  Hopefully the above SKE command will help
> things.  If not, let us know the exact SKE commands (or unix script
> commands) you are using so I can try them here.  Hope this helps.


Really, the problem is somehow more complicated than it appears at
first sight. All this is related to LDM use for distribution of
images to Regional sites. The LDM feed is done in real time, as new
images are ingested by the SDI, so IMGCOPY runs at the Regional site
every 1/2 hour; it is not an scheduled command in a mcIDAS session, 
it gets triggered for every new data feed off the LDM flow. 
So, it would really correspond to an IMGCOPY scheduled to run at 1/2 
hour intervals, but not precisely at the very same minute. That's how 
I got my results. 

What lies behind my questions is if IMGCOPY is supposed to be used
with these restrictions in mind, or a macro similar to the old MOVIT
should be re-written instead. (Then, why was it ever dropped?)
For example, why can't I keep more than one day on-line in a DataSet?. 
(The image for say, the second day at 12:00Z would always overwrite 
the first day's 12:00Z slot, as the formula does not supply any DAY 
based offset) ...or...
Why should the time at which is run, affect the behaviour of the
IMGCOPY command? (You mentioned the CYCLE being set off sequence 
because of this?)( See code where tinc is the CYCLE value 3.0 )
C
C If they are doing it by the cycle method, figure out destination
C
      if (cycles .ne. 0) then
         mcycle = nint(ftime(iadir(5))/tinc)
         thisposit=1+mod(mcycle,destimgcount)


Would greatly appreciate if experience from unidata is available,
that's why I included them in the recipients list.

Txs,
Pepo
+*+*+*+*+*+*+*+*+*+*+*+*+*+*+**+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*
Jose Juega                        Instituto Nacional de Meteorologia.
Tecnico de Sistemas                         Camino de las Moreras s/n
Area de Telematica      JJJJJJJ   JJJJJJJ   Ciudad Universitaria     
I.N.M.-Madrid-SPAIN       JJJ       JJJ     Aptdo. 285   28040 Madrid
                          JJJ       JJJ     Tel :     +34 91 581-9654
                      JJJ JJJ   JJJ JJJ     FAX :     +34 91 544-5307
UAM (McIDAS)           JJJJJ     JJJJJ    e-mail:         address@hidden
+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*++*++*+*+*+*+*+*+*+*+*+*+*+*+*+
Visit http://www.geocities.com/SiliconValley/5659