[evlatests] Excessive correlator configurations

[Rick Perley]

     The 27-hour 'flux density model' run is completed, and the data 
have been downloaded onto lustre.  In examining what is there, I have 
encountered a strange peculiarity:

     The run has 10 unique frequency tunings (one for each band, plus an 
extra for the X-band referenced pointing).  There should, therefore, be 
10 'correlator configurations' in the archive.

     In fact, there are 51.

     Each of the 10 unique frequency tunings has resulted in 5 to 6 
separate, but identical, 'configurations' in the database.  I've seen 
this behavior once before -- in one of the beam holography runs -- and 
the 'fix' was relatively easy:  'DBCON' correctly rejoins the disparate 
datasets.   There seems to be no loss of data.

     But it's a nuisance to have to load 51 separate files, then join 
them all together.

     What is happening here?

     Rick