Date: Fri, 11 Oct 2024 01:17:54 +0000 (UTC)
From: "Efrain J. Berdecia" <ejberdecia@yahoo.com>
Reply-To: "Efrain J. Berdecia" <ejberdecia@yahoo.com>
To: xof@thebuild.com, sud <suds1434@gmail.com>
Cc: pgsql-general <pgsql-general@lists.postgresql.org>
Message-ID: <2079480455.569271.1728609474771@mail.yahoo.com>
In-Reply-To: <56CE69DD-FFC8-45A4-8EF6-513BB011310F@thebuild.com>
References: <CAD=mzVWLN35FHPxN0Ajh2vnDkFLyg9f0vT6EKF9uPm7CYkHS5g@mail.gmail.com> <56CE69DD-FFC8-45A4-8EF6-513BB011310F@thebuild.com>
Subject: Re: Question on indexes
MIME-Version: 1.0
Content-Type: multipart/alternative; 
	boundary="----=_Part_569270_32428588.1728609474770"
Content-Length: 2545
Archived-At: <https://www.postgresql.org/message-id/2079480455.569271.1728609474771%40mail.yahoo.com>
Precedence: bulk

------=_Part_569270_32428588.1728609474770
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: quoted-printable

They are extremely efficient for joins!!!

Yahoo Mail: Search, Organize, Conquer=20
=20
  On Thu, Oct 10, 2024 at 2:52 PM, Christophe Pettus<xof@thebuild.com> wrot=
e:  =20

> On Oct 10, 2024, at 11:49, sud <suds1434@gmail.com> wrote:
>=20
> Hi,
> I have never used any 'hash index' but saw documents in the past suggesti=
ng issues around hash index , like WAL doesnt generate for "hash index" whi=
ch means we can't get the hash index back after crash also they are not app=
lied to replicas etc.

That's very old information.=C2=A0 Hash indexes are correctly WAL-logged si=
nce (IIRC) version 10.

 =20

------=_Part_569270_32428588.1728609474770
Content-Type: text/html; charset=UTF-8
Content-Transfer-Encoding: quoted-printable

They are extremely efficient for joins!!!<br><br><div id=3D"ymail_android_s=
ignature"><a id=3D"ymail_android_signature_link" href=3D"https://mail.oneli=
nk.me/107872968?pid=3DNativePlacement&amp;c=3DGlobal_Acquisition_YMktg_315_=
EmailSignatureGrowth_YahooMail:Search,Organize,Conquer&amp;af_sub1=3DAcquis=
ition&amp;af_sub2=3DGlobal_YMktg&amp;af_sub3=3D&amp;af_sub4=3D100000945&amp=
;af_sub5=3DOrganizeConquer__Static_">Yahoo Mail: Search, Organize, Conquer<=
/a></div> <br> <blockquote style=3D"margin: 0 0 20px 0;"> <div style=3D"fon=
t-family:Roboto, sans-serif; color:#6D00F6;"> <div>On Thu, Oct 10, 2024 at =
2:52 PM, Christophe Pettus</div><div>&lt;xof@thebuild.com&gt; wrote:</div> =
</div> <div style=3D"padding: 10px 0 0 20px; margin: 10px 0 0 0; border-lef=
t: 1px solid #6D00F6;"> <br clear=3D"none"><div class=3D"yqt9100655869" id=
=3D"yqtfd74638"><br clear=3D"none">&gt; On Oct 10, 2024, at 11:49, sud &lt;=
<a shape=3D"rect" ymailto=3D"mailto:suds1434@gmail.com" href=3D"mailto:suds=
1434@gmail.com">suds1434@gmail.com</a>&gt; wrote:<br clear=3D"none">&gt; <b=
r clear=3D"none">&gt; Hi,<br clear=3D"none">&gt; I have never used any 'has=
h index' but saw documents in the past suggesting issues around hash index =
, like WAL doesnt generate for "hash index" which means we can't get the ha=
sh index back after crash also they are not applied to replicas etc.</div><=
br clear=3D"none"><br clear=3D"none">That's very old information.&nbsp; Has=
h indexes are correctly WAL-logged since (IIRC) version 10.<div class=3D"yq=
t9100655869" id=3D"yqtfd50130"><br clear=3D"none"><br clear=3D"none"></div>=
 </div> </blockquote>
------=_Part_569270_32428588.1728609474770--