统计对象大小信息的函数和子查询的Bug(二)

2014-11-24 03:31:09 · 作者: · 浏览: 8
ERY PLAN --------------------------------------------------------------------------------------------------------------------------------- Gather Motion 64:1 (slice7; segments: 64) (cost=445.41..10096.03 rows=1 width=128) -> Hash Left Join (cost=445.41..10096.03 rows=1 width=128) Hash Cond: c.reltablespace = t.oid -> Redistribute Motion 64:64 (slice5; segments: 64) (cost=443.06..10092.81 rows=1 width=132) Hash Key: c.reltablespace -> Hash Left Join (cost=443.06..10092.22 rows=1 width=132) Hash Cond: c.relnamespace = n.oid -> Redistribute Motion 64:64 (slice3; segments: 64) (cost=54.53..9703.24 rows=1 width=72) Hash Key: c.relnamespace -> Hash EXISTS Join (cost=54.53..9702.65 rows=1 width=72) Hash Cond: c.relname = b.tablename::name -> Redistribute Motion 1:64 (slice1) (cost=0.00..9621.26 rows=10570 width=72) Hash Key: c.relname -> Seq Scan on pg_class c (cost=0.00..2751.39 rows=676419 width=72) Filter: relkind = 'r'::"char" -> Hash (cost=53.49..53.49 rows=2 width=24) -> Redistribute Motion 64:64 (slice2; segments: 64) (cost=0.00..53.49 rows=2 width=24) Hash Key: b.tablename::name -> Seq Scan on t b (cost=0.00..51.83 rows=2 width=24) -> Hash (cost=388.10..388.10 rows=1 width=68) -> Redistribute Motion 1:64 (slice4) (cost=0.00..388.10 rows=35 width=68) Hash Key: n.oid -> Seq Scan on pg_namespace n (cost=0.00..365.35 rows=2240 width=68) -> Hash (cost=2.32..2.32 rows=1 width=4) -> Redistribute Motion 1:64 (slice6) (cost=0.00..2.32 rows=2 width=4) Hash Key: t.oid -> Seq Scan on pg_tablespace t (cost=0.00..1.02 rows=128 width=4) (27 rows)


该问题应该是个Bug,等待TSE给出Fix或者没有Fix而只能等到下个版本升级了.
-EOF-