What are the differences between using SQLite and Berkeley DB?
The programming interface (API) and SQL language supported are identical: Berkeley DB uses that part of SQLite without any changes. Similarly for the special commands understood by the command line tool. However, the different storage architecture (and time available for the release) has meant that there are some caveats about 100% compatibility: • some SQLite pragmas are unsupported or have different semantics and default values (e.g. cache size) • differences in locking semantics generally result in higher concurrency, but may also lead to deadlocks occurring in Berkeley DB that did not occur in SQLite (for example, if two transactions are executed concurrently, one updating table A then table B, and the other doing the updates in the opposite order).
