Re: COMMIT frequency

From: Tad Davis <davist_at_isc.upenn.edu>
Date: 1998/03/17
Message-ID: <350EE838.82D097BA@isc.upenn.edu>#1/1

Allan Speir wrote:

> The deciding factor here is performance. Normally committing aftr every
> update gives poorer performance than committing after each row.

In the type of program in question here, each row represents a single update. The routines I'm working with read a single table and update rows in that table on a case-by-case, row-by-row basis. There is no logical connection between rows. In that sense, a commit after every update is definitely cleaner.

Tad

-- 
Tad Davis                  Lead Programmer Analyst
davist_at_isc.upenn.edu       Information Systems and Computing
voice 215-898-7864         Administrative Information Technologies
fax   215-898-0386         University of Pennsylvania

Received on Tue Mar 17 1998 - 00:00:00 CST