Github user eychih commented on the issue:
    @ravipesala @jackylk It looks like the discussion become uninteresting and 
unproductive now.  I just try to make some constructive comments.  The comments 
I made in the previous discussions remain true.  In terms of MV, the 
performance bottleneck comes from the amount of data we need to process and how 
we process them, e.g. using cache, rather than matching time of the logical 
plan.  The reason I take the approach in chowbranch because the old code is 
difficult to understand and to extend.  This is not just my opinion.  I will 
stop here and not make any further comment.   


Reply via email to