[daip] DBCON on many files

[Eric Greisen]

On 04/04/2018 07:21 PM, Walter Brisken wrote:
> 
> Hi Eric,
> 
> DBAPP seems not to be working.  The error I get and the image headers 
> are shown below.  Note that I've tried various FQTOL values (including 
> -1), and tried multiple inseq / in2seq values from a string of 13 such 
> files on disk.  I've also experimented with outname/outclass/outseq 
> being an existing file or one that is to be created.  Same error message 
> occurs in any case.  The two data files in the example below can be 
> found in:
> 
>    /home/parallax/BB388/testfiles
> 
> Inputs and error message:
> 
>> inp
> AIPS 2: DBAPP     appends one or more uv data sets to the output file
> AIPS 2: Adverbs     Values                 Comments
> AIPS 2: ----------------------------------------------------------------
> AIPS 2: INNAME     'NEP0002'               Input UV file name (name)
> AIPS 2: INCLASS    'V0.1'                  Input UV file name (class)
> AIPS 2: INSEQ        10                    Input UV file name (seq. #)
> AIPS 2: INDISK        5                    Input UV file disk unit #
> AIPS 2: IN2SEQ       10                    Upper input file seq. #
> AIPS 2: OUTNAME    'NEP0002'               Output UV file name (name)
> AIPS 2: OUTCLASS   'V0.1'                  Output UV file name (class)
> AIPS 2: OUTSEQ        9                    Output UV file name (seq. #)
> AIPS 2: OUTDISK       5                    Output UV file disk unit #.
> AIPS 2: FQTOL       100                    > 0 tolerance to not renumber
> AIPS 2:                                      FQs of dataset 2 (in kHz)
> AIPS 2:                                    -1 => no renumbering.
>> go
> DBAPP2: Task DBAPP  (release of 31DEC18) begins
> DBAPP2: NUMBERS OF POINTS ON AXES DO NOT MATCH
> DBAPP2: ERROR AROSE WITH INPUT FILE SEQUENCE NUMBER    10
> DBAPP2: Purports to die of UNNATURAL causes
> DBAPP2: parallax     31DEC18 TST: Cpu=       0.0  Real=       0
> AIPS 2: Resumes
> 

i have found several issues:

1. UT1UTC changes between your 2 files - that was treated as a hard 
error.  Changed it to a warning.

2. The check on times did not include a time offset, so concluded that 
the output file came after the input.

3. But worse - the UV header Common changed some long time ago and the 
parallel versions of it in DBAPP were not updated.  This trashed the 
data and could (although not in the case I tested) lead to the error you 
got.

Will be fixed in 31DEC17 this afternoon, and 31DEC18 after 10 pm tonight.

Eric