Re: INSERT/UPDATEs cycles and lack of phantom locking - Mailing list pgsql-sql

From Florian Weimer
Subject Re: INSERT/UPDATEs cycles and lack of phantom locking
Date
Msg-id 827j291vo9.fsf@mid.bfk.de
Whole thread Raw
In response to Re: INSERT/UPDATEs cycles and lack of phantom locking  (Tom Lane <tgl@sss.pgh.pa.us>)
Responses Re: INSERT/UPDATEs cycles and lack of phantom locking
List pgsql-sql
* Tom Lane:

> Why do you think that?  If you're running in READ COMMITTED mode then
> each statement takes a new snapshot.

Ah, I should have explained that.  I might need the SERIALIZABLE
isolation level in the future (this code doesn't need it, but other
things in the same transaction might require it).

In addition, it occurred to me that I get the INSERT failure only if
there is a suitable PRIMARY KEY/UNIQUE constraint on the table.  I
haven't got that in all cases, so I need that advisory locking anyway,
I fear.

-- 
Florian Weimer                <fweimer@bfk.de>
BFK edv-consulting GmbH       http://www.bfk.de/
Durlacher Allee 47            tel: +49-721-96201-1
D-76131 Karlsruhe             fax: +49-721-96201-99


pgsql-sql by date:

Previous
From: "Aaron Bono"
Date:
Subject: Re: User Permission
Next
From: Richard Broersma Jr
Date:
Subject: Re: User Permission