Closed moggers87 closed 10 years ago
In [4]: watson.search("gnu").count()
Out[4]: 1680
In [5]: watson.search("mail").count()
Out[5]: 111338
I don't know why I didn't think of this before :)
Looking at the queries that Watson is doing, the only possible speed up would be to replace the subquery with a join. Not sure if that will help.
For your viewing: http://pamrel.lu/d7dae/
Query plan: http://pamrel.lu/329b1/
Query analysis: http://pamrel.lu/a2b71/
I'm closing this as #170 is now under way!
For example, "gnu" takes 0.8 seconds yet "mail" takes 50 seconds.
Both return 1 inbox and the maximum number of emails (10). It might be that either query slicing isn't working (bug upstream) or the SQL that's being generated is really really slow when applying LIMIT
Further investigation required.