[daip] Extension table missing {External}

[Eric Greisen]

You could try "ls -lt" in the data area.  I suspect that the files causing a problem are all rather old and that your files are not old.  The file names are ttDnnnvvv.uuu; where tt is the file type, D is the version code (not changed in many years), nnn is the catalog number, vvv is the version number (i.e. when you have multiple plot files), and uuu is the user number.  Note that nnn, vvv, and uuu are in what we call "extended hex" or base 36 (0-9, A-Z).  If you "source LOGIN.SH" in the $AIPS_ROOT area then the command REHEX xxx will convert xxx in extended hex into a regular number.  EHEX yyy converts yyy into extended hex.  Files that are not known to your CATALOG can be recovered if they have a CB file (the header).  If they don't then they should be deleted.  The best way to look may require a "ls -l ??Dnnn???.uuu\;" which will show you all files for the catalog number nnn.  Note that you have to do that 1 nnn at a time but the ls -lt may not get all the files for an nnn right together.

________________________________
From: mz <mgnahz at gmail.com>
Sent: Friday, April 24, 2026 4:33 AM
To: Eric Greisen <egreisen at nrao.edu>
Subject: Re: Extension table missing {External}

Dear Eric,

What should I do to get the tables loaded right now? How to clean the data area without destroying all the data?

Best regards,
Ming


On Thu, 23 Apr 2026 at 23:43, Eric Greisen <egreisen at nrao.edu<mailto:egreisen at nrao.edu>> wrote:
Thank you.  I will not need the data file since we solved the problem without it.  When things go wrong in AIPS, files get left in the data area sometimes.  If they are not attached to a current entry in the catalog then all they can do is cause problems.  The GN table issue is not a problem.  The TABINI message says that your modern AN file (14 columns) encountered an old format AN table (the 2 columns were added in 2009 !)  I gather that things are working now, congratulations.

Eric Greisen

________________________________
From: mz <mgnahz at gmail.com<mailto:mgnahz at gmail.com>>
Sent: Thursday, April 23, 2026 3:24 AM
To: Eric Greisen <egreisen at nrao.edu<mailto:egreisen at nrao.edu>>
Subject: Re: Extension table missing {External}

Dear Eric,

Thank you very much for your willing to help.

First, the lines before TABINI is here:
CUSP  > FITLD5: UV table spanned time:   0/01:00:00 -  0/06:23:56
CUSP  > FITLD5:    75793 vis. written
CUSP  > FITLD5: Warning: table type GN is of zero length
CUSP  > FITLD5: TABINI: KEY COUNTS     35     51 OR COLS     14     12 DON'T MAT
CUSP  >         CH
CUSP  > FITLD5: TABINI: ERROR ON TYPE AN DISK  1 CNO      2 VERSION     1
It seems the GN table may have caused the cascade effect below.

Second, I have a public NAS server for you to access the FITS data:
https://kuohaouu.synology.me:55001
username: vlbi
password: Shao64386191
After you log in, directly go to "File Station" on the desktop and then click "vlbi-data" in the left panel, you will find a "pa002b" directory. Expand the "pa002b" directory, you will find my "pa002b.FITS" data. Its size is about 18GB.

I have attached the PRTTP output as well for your investigation in the attachment. As the orphan files in the data area, I did encounter some messages saying that CBD* UVD* SCD* HID* FQD* etc already exist. I deleted them in DATA directory and the FITLD task could go forward then.  Thank you very much for your consideration.

Best regards,
Ming



On Thu, 23 Apr 2026 at 07:16, Eric Greisen <egreisen at nrao.edu<mailto:egreisen at nrao.edu>> wrote:
I may be able to tell something from the answers to my previous email.  A full test would require your making the FITS file available to me.  If it is not too large, putting it on a public ftp site should work.

Eric
________________________________
From: mz <mgnahz at gmail.com<mailto:mgnahz at gmail.com>>
Sent: Wednesday, April 22, 2026 12:29 AM
To: Eric Greisen <egreisen at nrao.edu<mailto:egreisen at nrao.edu>>
Subject: Extension table missing {External}

You don't often get email from mgnahz at gmail.com<mailto:mgnahz at gmail.com>. Learn why this is important<https://aka.ms/LearnAboutSenderIdentification>
Dear Eric,

I've got the correlated FITS file from our correlator and when I load it into AIPS I got the ERROR messages like:
CUSP  > FITLD5: TABINI: ERROR ON TYPE AN DISK  1 CNO      2 VERSION     1
CUSP  > FITLD5: ANINI: ERROR   1 FROM TABINI OPCODE = WRIT
CUSP  > FITLD5: ANINI: ERROR   1 INITIALIZING ANTENNA TABLE FOR WRIT
CUSP  > FITLD5: AG2AN: ERROR   1 OPENING AN TABLE #    1
CUSP  > FITLD5: BT2AIP: ERROR   1 FORMING ANTENNA TABLE
CUSP  > FITLD5: ERROR      1 RETURNED FROM BT2AIP
CUSP  > FITLD5: Purports to die of UNNATURAL causes

I checked the loaded uv data that the AN and SU tables are missing. Is that a garbled file from the correlator, or is there any way to reconstruct those tables in AIPS? Thank you very much for your suggestion.

Best regards,
Ming Zhang


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listmgr.nrao.edu/pipermail/daip/attachments/20260424/22627913/attachment.html>