[daip] New Ticket - [!GIO-752054]: fitld DOCONCAT=1 says FILE ALREADY EXISTS

[Jan Wagner]

New Ticket: fitld DOCONCAT=1 says FILE ALREADY EXISTS

Hi! Two issues. One documentation issue(?) with FITLD. Another issue with DBCON.

I have older VLBA data that is split into two uvfits files because of too many IFs. To combine both files, I attempted two FITLD calls with doconcat=1, relying on the HELP "If DOCONCAT is TRUE then FITLD will look over all files matching the OUTNAME, OUTCLAS, OUTSEQ, OUTDISK parameters to try to find one that it can append data to." 

Either it is a documentation bug or I'm misunderstanding the help, but, appending does not work. Commands are below:

> task 'fitld'; default;
> outname='BB227APC';outclass='UVDATA';outseq=1;outdisk=1;optype='uv';ncount=1;nfiles=0;
> clint=0.1;digicor=1;doconcat=1;
> datain='/Exps/MCP/BB227/BB227A_P.uvfits_1
> go  $ loads the first part of data just fine

Now trying to append:

> datain='/Exps/MCP/BB227/BB227A_P.uvfits_2
> go  $ fails to append to existing file

Results in:

FITLD1 11:02:18 Task FITLD  (release of 31DEC11) begins
FITLD1 11:02:18 Found MULTI    observed on 10-DEC-2006
FITLD1 11:02:18 FILE ALREADY EXISTS
FITLD1 11:02:18 ERROR     -1 RETURNED FROM UVFHDR
FITLD1 11:02:18 Purports to die of UNNATURAL causes

Another approach: setting ncount=2 and having datain end with '_'. With outseq=1 set, FITLD fails on the second file. With outseq=0 set, FITLD creates two catalog entries, again not a single concatenated entry.

Last approach: importing two files separately. Then performing a DBCON. 

This is my second issue. The DBCON fails with an IF AXIS HAS UNEQUAL ATTRIBUTES.

The uvfits file are pathological:

UV data: LL and RR only, no cross-pol data
File A: RL,L,R: RL freq 1, L freq 2, R freq 3 
File B: L,RL,L: L freq 3, RL freq 4, L freq 5
Imhead on file 1: 2-stokes 3-IF
Imhead on file 2: 1-stokes 4-IF

There seems to be a failure in AIPS (31DEC11) to recognize the middle 'LR' pair in File B as 2-stokes. Trying a VLBAFPOL on either file simply claims that labelling is already correct.

I expect DBCON to be able to merge the two files. Alternatively FITLD should recognize File B middle 'LR' pair by its identical frequencies as indeed representing 1 IF with 2 Stokes, instead of 2 IFs with 1 Stokes.

The expected result of DBCON would then be a 5 IF 2-Stokes UV dataset with frequency table: RL,L,RL,RL,L.

kind regards,
 - Jan


Ticket Details
===================
Ticket ID: GIO-752054
Department: AIPS Data Processing
Priority: Default
Status: Open
Link:  https://help.nrao.edu/staff/index.php?_m=tickets&_a=viewticket&ticketid=1352