[evlatests] 'Non-closing' missing data ...

[Rick Perley]

    Vivek alerted me to something in one of his databases -- I decided 
to look into this with the 2 GHz BW 6cm test data taken yesterday 
(Tuesday). 
   
    I previously reported that 3.6% of the data were reported as 
'integer zero', but that I had not yet identified where these were (this 
being because none of my initial listing showed any gaps or holes of any 
kind).  I also reported that antenna 1 had no data in subband 1.  These 
two facts are related ...

    It turns out that it is true that antenna 1 has no data in subband 
11.  All entries are blank for all channels.  But it's also true that 
many other antennas are *partially* missing in the subband.  And the 
missing bits don't close -- meaning that the absences are not uniquely 
antenna based.   A matrix listing shows the curious pattern of 
absences.  The only 'easy' statement is that the RR and LL gaps are 
identical.  Some antennas have nearly all their baselines, some have 
very few (and antenna 1 has none). 
   
I checked one other subband -- all matrix entries are complete.  There's 
something special about subband 11. 

    Also notable is the length of time it takes to read these data:  
each matrix listing (asking for a single record from a single channel at 
a single time, for all antennas from one subband) takes at least 15 
minutes wall clock time.