One transaction like you did is best. I recently ran a test which ran pretty well with a commit every 1M records. Doing every 100,000 records slowed things down dramatically.
-----Original Message----- From: sqlite-users-boun...@sqlite.org [mailto:sqlite-users-boun...@sqlite.org] On Behalf Of joe.fis...@tanguaylab.com Sent: Monday, December 31, 2012 12:32 PM To: sqlite-users@sqlite.org Subject: [sqlite] 1.1GB database - 7.8 million records Very impressive. With SQLite 3.7.14.1 Took 4 minutes to load a 1.5GB MySQL dump with 7.8 million records. Count(*) takes 5 seconds. Even runs on a USB key. Wow! Also loaded a smaller one (33MB database [30 tables/dumps] in 10 seconds, largest file had 200,000 records). I wrapped the 7.8 million records in one [BEGIN TRANSACTION;] [COMMIT TRANSACTION;] block. Had to use VIM to edit the file. Using the Transaction is significantly faster with a large number of inserts. What's the rule of thumb on how many records per transaction? Does it matter how many are used, is one transaction OK? Joe Fisher _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
