| From: | "Christopher Browne" <cbbrowne(at)gmail(dot)com> |
|---|---|
| To: | "Wright, George" <George(dot)Wright(at)infimatic(dot)com> |
| Cc: | pgsql-novice(at)postgresql(dot)org |
| Subject: | Re: function transaction scope question |
| Date: | 2009-01-14 15:47:54 |
| Message-ID: | d6d6637f0901140747t90b4d00q2b35af813d3ebc9@mail.gmail.com |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-novice |
On Wed, Jan 14, 2009 at 9:41 AM, Wright, George
<George(dot)Wright(at)infimatic(dot)com> wrote:
> If a function in PL/pgSQL is wrapped by a transaction, does that same single
> transaction encompass other functions defined elsewhere that this function
> calls?
There's a rather stronger condition than that to be claimed...
*EVERY* PostgreSQL stored function call (irrespective of language)
runs within the context of a particular transaction. That transaction
context will therefore include all of the following:
- any functions that the stored function may call
- any triggers that may be fired by virtue of table updates
- invocation of NOTIFY (but *not* the notifications that result in
listening connections)
That's quite comfortably stronger than what you appear to be hoping for :-).
--
http://linuxfinances.info/info/linuxdistributions.html
Steve Martin - "I like a woman with a head on her shoulders. I hate necks."
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Wright, George | 2009-01-14 19:56:10 | another transaction question |
| Previous Message | John DeSoi | 2009-01-14 14:59:14 | Re: function transaction scope question |