[daip] FITS files in FITAB vs FITTP

[Eric Greisen]

Shami Chatterjee writes:

 > I'm trying to export FITS files of autocorrelation data only for
 > downstream work with IDL), and it apears that with identical inputs
 > (and defaults) and the same data file, FITTP and FITAB produce very
 > different FITS files.
 > 
 > In particular, while IDL will read FITTP fits files, it does not like
 > files written by FITAB, primarily because FITAB labels the length of
 > NAXIS2 = 0.
 > 

        I am suprised that IDL can read uv data files of any sort.
The FITS files are very different.  FITAB is a now standard binary
table representation of the data so there is no primary data set.  The
uv data are in one of the tables.  FITTP uses random groups which has
(sort of) a primary data set.  The help file for FITAB warns (I
believe) that many other systems will not yet know how to read its
output.

Eric