Hi again, I've been experimenting with limiting memory usage in our SQLite-based app. Ran into an unrelated oddity that I thought I'd ask about:
We're running a couple of massive upgrade steps on over 5 million quite large (70+ columns) rows. There are two unrelated steps; 1) DROP COLUMN-replacement where all data is copied to a table with the new schema, the old table is dropped, and the new table is renamed to replace 2) UPDATE statement of all rows These are obviously huge transactions, so I was expecting to see the WAL file grow to about the same size as the original database. But for some reason, the WAL-index (-shm) file also grows to about 40MiB in size. From the docs, I've got the impression that it would typically stay at around 32KiB. Does this seem normal? I just ran into some high-level documentation for the WAL-index [1] where it indicates there's a page correlation -- is it growing so much because we're touching so many pages? SQLite v.3.14.1. Thanks, - Kim [1] https://www.sqlite.org/fileformat2.html#walindexformat _______________________________________________ sqlite-users mailing list sqlite-users@mailinglists.sqlite.org http://mailinglists.sqlite.org/cgi-bin/mailman/listinfo/sqlite-users
