>>32
Considering the way the BerkleyDB api is designed, this says a heck of a lot more about your code than it does BDB. The API only provides a "get" function for retrieving records, how you search the database is entirely up to you. But you already knew that, right?
It's particularly laughable when you consider the fact that BDB is used by many email servers, tel-com, commercial switches/routers that all regularly handle terabytes worth of data without so much as a hiccup.
For reference:
http://www.oracle.com/database/docs/Berkeley-DB-v-Relational.pdf
>By avoiding context switches and minimizing copies, Berkeley DB gets outstanding performance on commodity hardware. ... Read throughput, reading a single record at a time, was 466,623 records per second. Using a high performance bulk retrieval interface for large sequential scans of the database, the same system was able to read 13,501,800 records per second.
*sigh*
If you don't know what BDB is, or how to use it... then chances are you don't need it. SQLite is amazing, and good for almost everything. BDB is, however, not nonsense and extremely powerful if used correctly.