[HACKERS] [PATCH] Overestimated filter cost and its mitigation

n, only 84.4 tuples of pgbench_history are expected to be evaluated with SubPlan 1. It is close to actual number 94 to some extent, and total cost seems much more reasonable than cost with current upstream. Any thoughts? Regards, Yuto Hayamizu Mitigate-filter-cost-overestimation.patch Descri

[HACKERS] HEAD crashes with assertion and LWLOCK_STATS enabled

s entries for that locks. So I feel it needs some refinements by you hackers. From f96708d14ab0073abd95c463eaf8d60db42f411d Mon Sep 17 00:00:00 2001 From: Yuto Hayamizu Date: Tue, 20 May 2014 16:19:56 +0900 Subject: [PATCH] pre-populating lwlock_stats entries --- src/backend/storage/lmgr/