Hi, Please find attached my 4th version this patch.
This is now based on v17 of Amit's faster partition pruning patch [1]. It also now includes Beena's tests which I've done some mostly cosmetic changes to. I've also fixed a few bugs, one in a case where I was not properly handling zero matching partitions in nodeAppend.c. Another change I've made is to now perform the partition pruning at run-time using a new memory context that's reset each time we redetermine the matching partitions. This was required since we're calling a planner function which might not be too careful about pfreeing memory it allocates. A test case I was running before making this change ended out failing to palloc memory due to OOM. I've not done anything about reducing the cost of the Append path when runtime pruning is enabled. I'm still thinking over the best way to handle that. [1] https://www.postgresql.org/message-id/58c3e20a-a964-4fdb-4e7d-bd833e9be...@lab.ntt.co.jp -- David Rowley http://www.2ndQuadrant.com/ PostgreSQL Development, 24x7 Support, Training & Services
runtime_prune_drowley_v4.patch
Description: Binary data
