From: | Randolf Richardson <rr(at)8x(dot)ca> |
---|---|
To: | pgsql-general(at)postgresql(dot)org |
Subject: | Re: Large objects [BLOB] again - general howto |
Date: | 2003-11-26 04:29:51 |
Message-ID: | Xns943ECE71A8448rr8xca@200.46.204.72 |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-general |
[sNip]
> Here's a quick list of my experiences with BLOB's and such.
>
> Performance is just fine, I get about 1M hits a month and haven't had
> any problems. Use a BLOB if you don't need to search though the data.
> The main reason being that bytea and text types are parsed. To explain,
> your entire SQL statement has to be parsed to make sure it's valid, this
> includes you 100 meg jpeg. Slow, REAL slow.
It appears that BLOBs are very well justified then.
> I have found that it is best to have a separate connection for BLOB's
> and one for everything else. Mind you, this is with Java, but the
> autocommit settings on the connection don't appear to be thread safe, so
> in high traffic you can accidentally cut off a transfer, or stop one
> before it ever starts.
What is your reason for using a separate connection?
--
Randolf Richardson - rr(at)8x(dot)ca
Vancouver, British Columbia, Canada
Please do not eMail me directly when responding
to my postings in the newsgroups.
From | Date | Subject | |
---|---|---|---|
Next Message | Tom Lane | 2003-11-26 05:39:27 | Re: performance versus order of fields in row |
Previous Message | Randolf Richardson | 2003-11-26 04:27:22 | Re: [OT... But couldn't resist!] Master/Slave |