Re: count on transaction ID - Mailing list pgsql-performance

From Kevin Grittner
Subject Re: count on transaction ID
Date
Msg-id 4F5DC49D02000025000460F6@gw.wicourts.gov
Whole thread Raw
In response to count on transaction ID  (ddgs <gsmaths@gmail.com>)
Responses Re: count on transaction ID
List pgsql-performance
ddgs <gsmaths@gmail.com> wrote:

> It is a discussion about the transaction ID wraparound in
> PostgreSQL.

Hopefully you've seen this:

http://www.postgresql.org/docs/current/static/routine-vacuuming.html#VACUUM-FOR-WRAPAROUND

> However, what is the fundamental definition if transaction ID.

As the cited page states, it is 32 bits.  It is considered a
"circular" number space.

> select * from table where ID=1:10000
> it is consider as one transaction or 10000 transactions.

In SQL it would be a syntax error; it doesn't really mean anything.
And it seems that you may be confused about the difference between
transaction IDs, object IDs, and user-defined ID columns on tables.
What is the "ID=1:10000" syntax supposed to mean?

-Kevin

pgsql-performance by date:

Previous
From: "McGehee, Robert"
Date:
Subject: Feature Request: No pg_dump lock on unlogged tables
Next
From: ddgs
Date:
Subject: Re: count on transaction ID