[daip] how can I combine these data sets?

[Walter Brisken]

Additional information:

DBCON says the following when attempting to combine two stokes I data 
bases (concurrently observed -- the two IFs from a 2AD VLA correlator mode 
where the IFs A and D had different frequencies (and obviously also 
different polarizations).  FILLM separated the IFs upon loading.  Note 
that bad things happen when outver is set since the two files that are 
created have the same name...  Try to concatinate two Stokes I databases:

localh> DBCON1: Task DBCON  (release of 31DEC03) begins
localh> DBCON1: Scaling u,v,w of data set 2 by  9.9253362E-01
localh> DBCON1: Scaling u,v,w of data set 1 by  9.9210346E-01
localh> DBCON1: RA and DEC match to within 3 microarcseconds each
localh> DBCON1: --I'm not going to shift positions
localh> DBCON1: FREQUENCIES DIFFER BY TOO MUCH:  1.37416250E+09  1.45396250E+09
localh> DBCON1: MULTI-FREQUENCY DATA BASES INCOMPATABLE
localh> DBCON1: Purports to die of UNNATURAL causes
localh> DBCON1: parsnip      31DEC03 TST: Cpu=       0.0  Real=       0

The output is the same for doarray = 1 or -1 or for fqtol = -1 or 1e8 .
I should also note that if Stokes I was not created for each database, 
DBCON would complain:

localh> DBCON1: Task DBCON  (release of 31DEC03) begins
localh> DBCON1: Scaling u,v,w of data set 2 by  9.9210346E-01
localh> DBCON1: Scaling u,v,w of data set 1 by  9.9253362E-01
localh> DBCON1: RA and DEC match to within 3 microarcseconds each
localh> DBCON1: --I'm not going to shift positions
localh> DBCON1: STOKES RANGES DO NOT OVERLAP
localh> DBCON1: Purports to die of UNNATURAL causes


These are single source files that need to be combined.  VBGLU seems to 
need an NX file:

localh> VBGLU1: Task VBGLU  (release of 31DEC03) begins
localh> VBGLU1: Create J0538+2817  .VBGLU .   1 (UV)  on disk  2  cno   73
localh> VBGLU1: VLBGIN: NO NX TABLE ATTACHED TO FILE  1
localh> VBGLU1: Destroyed UV image file: catno=     73 disk= 2
localh> VBGLU1: Purports to die of UNNATURAL causes
localh> VBGLU1: parsnip      31DEC03 TST: Cpu=       0.3  Real=       3

So I try this on the original CH 0 and LINE files.  According to the 
EXPLAIN file for VBGLU, I need to choose the file with most visibilities 
as the first file.  This crashes:

localh> VBGLU1: Task VBGLU  (release of 31DEC03) begins
localh> VBGLU1: Create AR514-L2    .VBGLU .   1 (UV)  on disk  2  cno   73
localh> VBGLU1: File AR514-L2    .LINE  .   1 on disk   2 covers
localh> VBGLU1: time range   0/ 14 41 40 -   0/ 16  2 15 and has   248400 vis
localh> VBGLU1: File AR514-L1    .LINE  .   1 on disk   2 covers
localh> VBGLU1: time range   0/ 14 41 40 -   0/ 16  2 15 and has   246470 vis
localh> VBGLU1: ZABORS: signal 11 received
localh> VBGLU1: ABORT!

But if I swap the input files it works.  So I continue...  
Checking the combined data with TVFLG shows that only source 1 of IF 1 is 
there -- IF 2 seems in tact.  So I stop for tonight...



I should probably also mention that this is an aips release from mid June 
on my laptop.  I don't think VBGLU or DBCON have changed since then. 

On Wed, 13 Aug 2003, Eric Greisen wrote:

> What does DBCON say?
> 
> E
>