From: Tom Lane <tgl@sss.pgh.pa.us>
To: lpy.henu@gmail.com
cc: pgsql-bugs@lists.postgresql.org
Subject: Re: BUG #17017: Two versions of the same row of records are returned
 in one query
In-reply-to: <17017-c37dbbadb77cfde9@postgresql.org>
References: <17017-c37dbbadb77cfde9@postgresql.org>
Comments: In-reply-to PG Bug reporting form <noreply@postgresql.org>
	message dated "Tue, 18 May 2021 13:26:24 -0000"
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-ID: <3229566.1621354699.1@sss.pgh.pa.us>
Date: Tue, 18 May 2021 12:18:19 -0400
Message-ID: <3229567.1621354699@sss.pgh.pa.us>
Archived-At: 
 <https://www.postgresql.org/message-id/3229567.1621354699%40sss.pgh.pa.us>
Precedence: bulk

PG Bug reporting form <noreply@postgresql.org> writes:
> Schema and Initial data: 
> Create Table t(a int primary key, b int);
> Insert into t values(1,2);
> Insert into t values(2,3);

> [Time0, SessonA]
>> Begin;
>> set transaction isolation level repeatable read;
>> Select * from t where a=1;

> [Time1, SessonB]
>> Begin;
>> set transaction isolation level read committed;
>> Delete from t where a=2;
>> Commit;

> [Time2, SessonA]
>> Insert into t values(2,4);
>> Select * from t where a=2;

> Here, we expect PostgreSQL Server to return a row:
> 2 3

> However, it returns two rows:
> 2 4
> 2 3

In repeatable read mode, this doesn't surprise me.  Session A
surely must return the (2,3) row, since it isn't supposed to
"see" the results of Session B's commit yet.  And it would be
mighty surprising for it not to see its own insertion, so you
get (2,4) as well.

In serializable mode, I'd expect this situation to throw a
serialization error, and it does (but you must run BOTH
transactions in serializable mode; there are not guarantees
with a mix of serializable and non-serializable transactions).

So I think the answer is "if you care about this, use
serializable mode".  Repeatable read mode meets the
requirements of the SQL spec, but it doesn't guarantee
no anomalies.

			regards, tom lane