| From: | Joe Conway <mail(at)joeconway(dot)com> |
|---|---|
| To: | Mircea Sarbu <msarbu(at)agora(dot)ro> |
| Cc: | pgsql-general(at)postgresql(dot)org |
| Subject: | Re: Using TEXT columns for binary content |
| Date: | 2003-04-21 16:01:24 |
| Message-ID: | 3EA415D4.5090807@joeconway.com |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-general |
Mircea Sarbu wrote:
> There are some hidden drawbacks for this approach? (I don't need any further
> processing on the binary content)
>
There are people who have reported using this method successfully on the
Postgres lists. Only downsides I can think of are:
1) You have to encode to base64 on the way in (bytea has to be escaped
on INSERT/UPDATE too)
2) You have to decode to binary on the way out (depending on your
middleware, you can use a binary cursor for bytea and avoid the
decode on SELECT)
3) The on disk representation is about 33% bigger compared to bytea
4) TEXT datatype is subject to being multibyte encoded -- I don't think
this affects the characters in the base64 set, but I'm not sure there
won't be any undesireable interactions. If you are using other than
'C' locale, test carefully.
Joe
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Murthy Kambhampaty | 2003-04-21 16:05:12 | Re: stddev returns 0 when there is one row |
| Previous Message | Steve Wampler | 2003-04-21 15:45:12 | Re: Are we losing momentum? |