Thread: How well clustered is a table?

How well clustered is a table?

From
Jonathan Blitz
Date:
I was wondering if there is some indication of how well clustered a table is.
 
In other words, when a Cluster command is performed then a table would be 100% clustered.
As updates etc are made the table clowly loses its clustering.
Is there any indication as to how "bad" it is at any one point?

Re: How well clustered is a table?

From
"A. Kretschmer"
Date:
In response to Jonathan Blitz :
> I was wondering if there is some indication of how well clustered a table is.
>
> In other words, when a Cluster command is performed then a table would be 100%
> clustered.
> As updates etc are made the table clowly loses its clustering.
> Is there any indication as to how "bad" it is at any one point?

Sure, because PG needs this information. See system_view pg_stats, column
correlation. This ranges from -1 to +1. When the value is near -1 or +1,
an index scan on the column will be estimated to be cheaper than when it
is near zero, due to reduction of random access to the disk.

For instance:

test=# select correlation from pg_stats where tablename='foo' and attname='key';
 correlation
-------------
    0.112908
(1 Zeile)

Zeit: 1,409 ms
test=*# cluster foo using idx_key;
CLUSTER
Zeit: 728,885 ms
test=*# analyse foo;
ANALYZE
Zeit: 59,970 ms
test=*# select correlation from pg_stats where tablename='foo' and attname='key';
 correlation
-------------
           1
(1 Zeile)

Zeit: 1,505 ms


Andreas
--
Andreas Kretschmer
Kontakt:  Heynitz: 035242/47150,   D1: 0160/7141639 (mehr: -> Header)
GnuPG: 0x31720C99, 1006 CCB4 A326 1D42 6431  2EB0 389D 1DC2 3172 0C99