[daip] Split CSLGET error

[Eric Greisen]

Arnaud Collioud wrote:
>>>   21806   04:37:29.0   00:00:30.0          197           30            2            1   0.000000E+00   0.0000000000D+00
>>>   21807   04:37:29.0   00:00:30.0          197           31            2            1   0.000000E+00   0.0000000000D+00
>>> Any idea why I have no output for this source?
>>> (I also tried without flagging, but that changed nothing)
>>>
>> This is a pretty severe test of the new code with 3 subarrays mixed.
>>
>> Try PRTAB with inext 'cl' and rparm = 5,1,1,0.1 to look only at subarray 1 lines.  The error message says there are none and the lines you sent such as the 2 above show subarray 2 only.
> 
> 
> With the param: rparm = 3 1 197 0.1 5 1 1 0.1, there is no resulting line. So for this source, only the subarray 2 has corrections
> (with only rparm = 5 1 1 0.1, I get *lots* of results for other sources).
> Do I need to remove the index lines corresponding to subarrays other than subarray 2? I thought that SPLIT was able to loop on subarrays and copy data only from subarrays with corrections...

I don't know how this error can arise.  CSLGET only reads record numbers 
that it is given in an array and those numbers are set by a routine that 
reads the CL file to begin with and which knows that maximum number.

the fundamental source of the failure is that you have UV data for this 
source in subarray 1 but no CL table entries for them.  That is a 
situation which is assumed never to arise and which SPLIT is not assumed 
to be able to surmount.

It would be very amusing to run the debugger in this situation to see 
what is happening where but I fear that your data set is rather large 
and so it would be hard to get it here for me to debug.  Perhaps it is 
not so bad.  A FITS file in an anonymous ftp area is how one would do 
this if it is not too terrible.

Eric Greisen