Hi,
It's a prepared sql statement on a non-partitioned table , 16millions
tuples and multiple indexes on this table. pk_x primary
key (aid,bid,btype) all 3 cols are bigint datatype, there is another index
idx_x(starttime,endtime) , both cols are "timestamp(0) without time
zone".
the
On Tue, 6 Dec 2022 at 20:17, James Pang (chaolpan) wrote:
> Could you provide the function name for generic plan selectivity estimation?
If you look at eqsel_internal(), you'll see there are two functions
that it'll call var_eq_const() for Consts and otherwise
var_eq_non_const(). It'll take the
al Message-
From: David Rowley
Sent: Tuesday, December 6, 2022 1:59 PM
To: James Pang (chaolpan)
Cc: pgsql-performance@lists.postgresql.org; jamespang...@gmail.com
Subject: Re: wrong rows and cost estimation when generic plan
On Tue, 6 Dec 2022 at 18:28, James Pang (chaolpan) wrote:
>->
On Tue, 6 Dec 2022 at 18:28, James Pang (chaolpan) wrote:
>-> Index Scan using idx_x_time on x (cost=0.44..8.48 rows=1
> width=2923) (actual time=8136.242..8136.242 rows=0 loops=1)
> Index Cond: ((starttime = $7) AND (endtime = $8))
> Filter: ((password IS NULL) AN
Hi,
It's a prepared sql statement on a non-partitioned table , 16millions tuples
and multiple indexes on this table. pk_x primary key (aid,bid,btype) all
3 cols are bigint datatype, there is another index idx_x(starttime,endtime)
, both cols are "timestamp(0) without time zone".
Hi,
It's a prepared sql statement on a non-partitioned table , 16millions tuples
and multiple indexes on this table. pk_x primary key (aid,bid,btype) all
3 cols are bigint datatype, there is another index idx_x(starttime,endtime)
, both cols are "timestamp(0) without time zone".
