On Thu, Apr 30, 2015 at 6:20 AM, Ciano Saraiva <saraiva.ufc@gmail.com>
wrote:
> Hello Postgres team,
>
> Wanted to give a suggestion in relation to [set enable_seqscan =3D off]
> command, the problem is that even disabled the SGBD can still use the
> sequential search depending on the query, my suggestion is that invez thi=
s
> command, it would be more intuitive even for IHC issues it was a command
> type [set enable_indexscan =3D ON ], this command does not sequential wou=
ld
> disable the search, but would enable the search using indixes, it is clea=
r
> that the SGBD that would decide which search would be better to use (
> sequential search and search using indexes). Because the command that is
> currently (set enable_seqscan =3D off) makes the user think it will turn =
off once
> and for all the sequential searches, but that's not what happens that he =
can
> still use.
>
>
=E2=80=8BI don't see anything about these so-called planner hints changing =
no
matter how potentially confusing they may be.
They are documented (and easily searched for) here:
http://www.postgresql.org/docs/9.4/static/runtime-config-query.html
and the point you are making is explicitly =E2=80=8Bexplained in the commen=
t about
enable_seqscan.
David J.
=E2=80=8BP.S. What are "SGDB", "IHC issues" and "invez"?