[daip] SPLAT vs. single-source multi-source UV databases

[Eric Greisen]

Walter Brisken writes:
 > 
 > Hello DAIP,
 > 
 > Is it legal to have a multi-source UV database containing only one source? 
 > It appears that SPLAT won't accomodate this ironic request.  The reason 
 > this is somewhat annoying is that in my pulsar pipeline there is sometimes 
 > one and sometimes more than one source being SPLAT out of an original UV 
 > database.  According to the SPLAT INP page, APARM(7)=0 will produce a 
 > multi-source database, but in practice it does not in the case that a 
 > single SOURCE is selected.  This has at least three implications for my 
 > pipeline: 1. the output name, which becomes SOURCE(1) rather than OUTNAME; 
 > 2. the propagation of certain tables (IM, CQ, IN, SU, ...), some of which 
 > I expect downstream; and 3. the application of CL tables copied from the 
 > in-beam calibrator.  For now I can work around this by using SOURCE='*'', 
 > but this is not really a great compromise.  Production of a true 
 > multi-source UV database containing the single source would be better. 
 > Equally good would be a true multi-source file but with no visibilities 
 > except for the requested source (ie maintain the SU table entries for all 
 > sources).  Perhaps the best comprimise would be for APARM(7)<0 to force 
 > multi-source even for a single requested source.
 > 
 > Thanks!
 > 
 > -W
 > 
 > (BTW -- thanks for the CQ table propagation that was recently added.  So 
 > far it looks good.  My pipeline now relies on this.)

I have changed SPLAT to honor the multi-source output request even
when only one source is selected.  The MNJ should fix this tonight for
use on Monday.

ERic Greisen