"Nested Loop (cost=486.21..84934.42 rows=25028 width=16) (actual time=54.627..328.019 rows=146523 loops=1)" " Buffers:

Question

"Nested Loop (cost=486.21..84934.42 rows=25028 width=16) (actual time=54.627..328.019 rows=146523 loops=1)" " Buffers:

shared hit=133393"

" -> Bitmap Heap Scan on task_type (cost=4.31..7.44 rows=2 width=16) (actual time=0.013..0.016 rows=2 loops=1)"

"        Recheck Cond: (process_id = ANY ('{2c7b0bfd-560c-479e-850a-a6128e562929,2c7b0bfd-560c-479e-850a-a6128e563030}'::uuid[]))"

"        Heap Blocks: exact=1"

"        Buffers: shared hit=3"

"        -> Bitmap Index Scan on task_type_process_id_idx (cost=0.00..4.30 rows=2 width=0) (actual time=0.007..0.007 rows=2 loops=1)"

"              Index Cond: (process_id = ANY ('{2c7b0bfd-560c-479e-850a-a6128e562929,2c7b0bfd-560c-479e-850a-a6128e563030}'::uuid[]))"

"              Buffers: shared hit=2"

" -> Bitmap Heap Scan on task (cost=481.91..42217.16 rows=24633 width=32) (actual time=31.059..146.706 rows=73262 loops=2)"

"        Recheck Cond: (type_id = task_type.id)"

"        Heap Blocks: exact=133029"

"        Buffers: shared hit=133390"

"        -> Bitmap Index Scan on task_type_id_idx (cost=0.00..475.75 rows=24633 width=0) (actual time=15.424..15.424 rows=73262 loops=2)"

"              Index Cond: (type_id = task_type.id)"

"              Buffers: shared hit=361"

"Planning Time: 0.273 ms"

"Execution Time: 334.484 ms"

explain (ANALYZE, BUFFERS) SELECT status_idFROM tfcs.task task WHERE                   (type_id in (select id from tfcs.task_type where process_id in ( '2c7b0bfd-560c-479e-850a-a6128e562929', '2c7b0bfd-560c-479e-850a-a6128e563030')))

План + квери.
Ещё напрягает, что используется 1.2 Гб буфера. При том, что для фильтров достаточно полей id, а он видимо тащит все поля у каждой записи?