From: | Oliver Jowett <oliver(at)opencloud(dot)com> |
---|---|
To: | Alvaro Herrera <alvherre(at)dcc(dot)uchile(dot)cl> |
Cc: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>, pgsql-hackers(at)postgresql(dot)org |
Subject: | Re: Nested Transaction TODO list |
Date: | 2004-07-05 05:15:26 |
Message-ID: | 40E8E3EE.9090207@opencloud.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
Alvaro Herrera wrote:
>>One generalization of this to nested transactions would be:
>>
>> SUBBEGIN [transactionname]
>> SUBCOMMIT [transactionname]
>> SUBABORT [transactionname]
>
>
> The only departure from the SAVEPOINT syntax is that you are able to
> "subcommit" a savepoint. Not sure how useful that is ...
One thing SUBCOMMIT [name] does allow is discarding savepoints / named
txns without rolling back their changes. That might be useful if we
allow nesting of names, e.g.:
SAVEPOINT save1
-- do work #1
SAVEPOINT save1 -- hides the earlier SAVEPOINT
-- do work #2
SAVEPOINT save2
-- do work #3
SUBCOMMIT save1 -- provisionally commits #2 and #3 to enclosing txn
-- do work #4
ROLLBACK TO save1 -- rolls back #1, #2, #3, #4
Other than that.. I assume we want SUBBEGIN/SUBCOMMIT/SUBABORT
regardless of how we do savepoints. Since savepoints are a subset of
what you can do with nested transactions, it seems appropriate that
SUBBEGIN etc. can do everything that savepoints can -- i.e. naming of
savepoints/transactions. And then SUBCOMMIT [name] is just there for
completeness.
-O
From | Date | Subject | |
---|---|---|---|
Next Message | Mario Weilguni | 2004-07-05 06:09:39 | Re: LinuxTag wrapup |
Previous Message | Tom Lane | 2004-07-05 04:41:08 | Re: Adding column comment to information_schema.columns |