[daip] Split CSLGET error

[Eric Greisen]

Arnaud Collioud wrote:
> Dear Eric,
> 
> I think there seems to be problems with (recent?) modifications of SPLIT.
> A the end of 2012 (exactly Nov 28, with 31DEC13), I splitted a multi-source uv-data and it produced 239 split files. Today, with an up-to-date 31DEC13, I only got 207 split files with the same uv file! Obviously, I made no changes in the uv-data file.
> 
> Did you apply some changes in SPLIT during this time that could lead to this kind of behaviour?
> 
> During the splitting, I have more errors of this kind:
> 
>>    1    2   06-JUN-2013  17:03:48     SPLIT     Create 2002-375    .SPLIT .   6 (UV)  on disk  1  cno  549
>>    1    6   06-JUN-2013  17:03:49     SPLIT     SPLCOP FOUND NO SAMPLES OUT OF   19737 POSSIBLE
>>    1    6   06-JUN-2013  17:03:49     SPLIT            IS SUBARRAY  1 IN CL/SN TABLE?
>>    1    6   06-JUN-2013  17:03:52     SPLIT                Previously flagged   flagged by gain         kept
>>    1    6   06-JUN-2013  17:03:52     SPLIT     Partially               0                   0              0
>>    1    6   06-JUN-2013  17:03:52     SPLIT     Fully                   0                   0              0
>>    1    8   06-JUN-2013  17:03:52     SPLIT     SPLCOP: ERROR - NO DATA WRITTEN
>>    1    4   06-JUN-2013  17:03:52     SPLIT     Destroyed  1 extension files of type FQ
>>    1    4   06-JUN-2013  17:03:52     SPLIT     Destroyed  1 extension files of type NX
>>    1    4   06-JUN-2013  17:03:52     SPLIT     Destroyed UV image file: catno=    549 disk= 1
>>    1    6   06-JUN-2013  17:03:52     SPLIT     PROBLEM WITH SOURCE: 2002-375         NO VIS SELECTED
> 
> and errors as I described in the previous email:
> 
>>    1    2   06-JUN-2013  16:41:36     SPLIT     Create 1313-333    .SPLIT .   6 (UV)  on disk  1  cno  365
>>    1    8   06-JUN-2013  16:41:37     SPLIT     TABIO: BAD LRNO=      627060 LIMIT      627059
>>    1    8   06-JUN-2013  16:41:37     SPLIT     CSLGET: ERROR   2 READING CALIBRATION TABLE
>>    1    4   06-JUN-2013  16:41:37     SPLIT     Destroyed  1 extension files of type NX
>>    1    4   06-JUN-2013  16:41:37     SPLIT     Destroyed UV image file: catno=    365 disk= 1
>>    1    6   06-JUN-2013  16:41:37     SPLIT     PROBLEM WITH SOURCE: 1313-333

Your answer confuses me - Nov 28 can only be Nov 28, 2012 (or earlier 
years) and they have nothing to do with aips version 31DEC13 which began 
in December 2012.  I am assuming that you are now running 31DEC13.  BUT 
and this is very important - have you run the MNJ to make sure that your 
current version of 31DEC13 is in fact current with ours?  I asked you to 
do that and you have not said that you did.

Yes - I have changed the calibration application routines so that they 
will not apply the calibration for source X to source Y.  This has 
exposed numerous difficulties in how the calibration application was 
previously coded.  One attribute of this - if your CL table lacks 
calibrations for some sources, then they will now fail as you show 
above, while previously CL values for some other source might be used 
and so the source would seem to have worked.  Note that CL table entries 
often contain values which depend on source (elevation gain corrections, 
source position and antenna position corrections, etc) - so it is 
incorrect to apply the CL values for one source to another.

I have another user who is claiming difficulties different from yours 
but indicating that there may still be problems with this code.  Your 
difficulties however suggest that you do not have the latest version.

Eric Greisen