On 2017-03-02 09:55, Christian Gollwitzer wrote: > you could do it "the (theoretical) right way", which is using > another table. Insert your keys into a table, maybe temporary one, > and then do > > select * from mumble where key in (select key from keytable) > > In theory that should also be faster, but unless you have a > thousand keys, there will be no noticeable difference.
It would depend on the frequency with which the same query is issued. Bulk inserts, followed by the query, followed by cleaning up the data would almost certainly be slower. Bulk inserting the data and then issuing thousands of queries against it could prove much faster (though as with all things performance-related, test, don't guess). But if the set of criteria changes frequently, the overhead of inserting/deleting the data will almost certainly swamp any potential speed-up. -tkc -- https://mail.python.org/mailman/listinfo/python-list