Chris,
I'll try and summarise.
1. We have a 60GB database collecting data. This database is accessed by
a single process once every five mins and around 5MB of data (approx
600-800 rows) is added. Data has never been deleted.
2. The database is getting too big for the server it's hosted on.
I've been following this thread with interest, but this just doesn't make
sense...
> Logically speaking SQLite shouldn't notice the difference in row order,
but things do slow down,
> even with analyse.
Are you accessing each row via its ID? Even so, that should still be
indexed.
I thought you
Dear all,
We think we have now found the issue with the slow commits.
We believe this is due to an inherent (and old) defect in our database
design. We think our original design has an implicit ordering of rows in
a table, when the table is only increasing this flaw in the design isn't
appare
3 matches
Mail list logo