Home/ Questions/Q 6376719
In Process

The Archive Base Latest Questions

Editorial Team
  • 0
Asked: 2026-05-25T01:46:51+00:00

I have a SQLite database that has pretty intensive repeated reads and occasional writes.

  • 0

I have a SQLite database that has pretty intensive repeated reads and occasional writes. However, the writes (because of indexing) tend to block the reads. I would like to read the on-disk database into a in-memory database and then have a way of syncing back to the on-disk when the machine is completely idle for maybe 5-10 seconds. I was briefly tempted to copy the tables from an attached on-disk database to an in-memory database, but it seems there should be superior way. I also considered transactions which are committed when the machine is idle (but will this block the intensive reads). The reads include the tables to be updated (or inserted), but the writes are not time-sensitive.

Share

Leave an answer

You must login to add an answer.

Need An Account,

1 Answer

  1. Editorial Team

    You should upgrade to SQLite 3.7.0 or later which includes Write Ahead Logging. This new method of locking allows reads while writing.

    http://www.sqlite.org/draft/wal.html

    To copy between an in-memory database and an on-disk database, you can use the backup API but it’s not exposed through the .NET wrapper yet.

    Also, by increasing your cache-size you can get the same performance from an on-disk database as an in-memory database–the whole thing can be cached in memory.

    Another option is using Oracle’s new version of BerkleyDB which has a SQLite front end including the same .NET wrapper and is a drop-in replacement for the official SQLite releases. They changed the locking mechanism to support page level locks instead of database level locks and greatly improved concurrency and therefore multi-connection performance. I haven’t used it myself, but I read good things.

    http://www.oracle.com/technetwork/database/berkeleydb/overview/index.html

    • 0