On Apr 8, 2012, at 12:25 AM, Josh Gibbs wrote:
> SQLite: INSERT OR IGNORE INTO test_table VALUES('prop','val');
> Postgres: INSERT INTO test_table SELECT 'prop','val' WHERE NOT EXISTS (SELECT
> 1 FROM test_table WHERE property='prop');
My 2¢…
The first variant, insert or ignore, is arguably mo
Thanks, but I can't do that because I'm batching up multiple writes in
transactions to get performance. The errors cause the whole transaction
to need to be rolled back.
On 8/04/2012 11:20 a.m., Igor Tandetnik wrote:
Josh Gibbs wrote:
The method that must be used is as follows:
CREATE TABL
Josh Gibbs wrote:
> The method that must be used is as follows:
>
> CREATE TABLE test_table (property TEXT PRIMARY KEY, value TEXT);
>
> SQLite: INSERT OR IGNORE INTO test_table VALUES('prop','val');
> Postgres: INSERT INTO test_table SELECT 'prop','val' WHERE NOT EXISTS
> (SELECT 1 FROM test_ta
Looking for some performance advice before I go testing this myself.
I'm porting some code that's currently running with SQLite as its DB
engine over to postgres. SQLite will still be an option so I need to
maintain compatibility across both engines. I've run into the common
postgres problem
4 matches
Mail list logo