On 2014-03-17 Mon 14:46 PM |, Viktor Dukhovni wrote: > > 20k records may seem like a lot of entries, but it is comparatively > a tiny Berkeley DB table. You've no reason to care about optimizing > it.
Thanks. > If you don't plan incremental updates, I'd like to suggest > CDB, because the on-disk format and API are more stable, and it > requires less memory. > >From the earlier archive links, it looked like the way to go. This default OS/arch & binary postfix package has btree support, but not CDB. If the tables get large, I may investigate this further.
