Re: Best Coding Practices
Rich Shepard <[email protected]>
| Newsgroups | gmane.comp.python.db.pysqlite.user |
|---|---|
| Message-ID | <[email protected]> |
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On Wed, 25 Oct 2006, Gerhard Häring wrote: > You should commit whenever you reached a consistent state of your data - > this rule is obviously only relevant when there are different entities in > your database (different tables). Gerhard, We do have about a dozen tables in the database. > Doing COMMITs when they're not really necessary will decrease performance > as each COMMIT implies a sync operation of the OS, so you're waiting on > buffers being flushed to disk here. This is important if you're doing mass > operations with loads of data. The usual procedure here is to commit after > every few hundred or thousand records that you insert during bulk loads. For the most part, the application is user-I/O bound. There is a very large amount of data to be entered for each model, and it's not a bulk insert but item-by-item. So issuing a COMMIT in our situation will keep the memory buffers flushed to disk. The actual run of a model is only a minute or two (for most situations), and the only table being written to is the log. Many thanks for the insight, Rich - -- Richard B. Shepard, Ph.D. | The Environmental Permitting Applied Ecosystem Services, Inc.(TM) | Accelerator <http://www.appl-ecosys.com> Voice: 503-667-4517 Fax: 503-667-8863 -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.5 (GNU/Linux) iD8DBQFFP74XugEIjC001L0RAjTBAJ4nEmwfPi1jY7NqVwRjWPPFhWS/+ACdGYkX J/+539M5+J3JC+g2n1J+drw= =Tz6Y -----END PGP SIGNATURE----- _______________________________________________ pysqlite mailing list pysqlite-IAPFreCvJWPBWskQ1e/[email protected] http://lists.initd.org/mailman/listinfo/pysqlite