HT-3026: Nightly etas overlap updates #97
Conversation
Caches cluster level counts. Could still be done for MPMs.
Also ensures orgs w/holdings in a cluster and no matches get copy_counts
jsteverman
force-pushed
the
HT-3026-nightly-etas-update
branch
from
0d715a7 to
b4275f1
Compare
jsteverman
changed the title
jsteverman
changed the title
|
Probably should be rebased, maybe should be squashed some. |
There was a problem hiding this comment.
Looks good to me. I like that this moves some of the Mongoid-specific stuff out of the overlap computation; given that we have to fetch the entire cluster anyway, this seems like this minimizes the number of times we need to look through the entire set of holdings for a cluster.
It may be worth spending a little bit of time looking at performance vs. size of the cluster to make sure that there isn't still an n^2 issue. @mwarin and I set up a test case with synthetic data to do something similar with holdings updates, which was helpful to verify that our changes were actually making an improvement; it was also helpful for being able to run the profiler without getting extremely noisy data. I think both those things would be helpful to do here as well, but shouldn't hold up merging this.
Caches cluster level counts. Could still be done for MPMs.