Hi Ken I found that performance began to fail miserably as the blob size increased above 500 bytes (has anyone else experienced this problem?). When I posted the problem, I think that someone stated the docs show the reasonable limit on blobs is 230 bytes. Anyways, I had to store the blobs direct to disk and use SQLite to track the addresses (no different than Oracle, MSSQL, MySQL etc).
Please notify if you can get your system to work! At 01:21 PM 9/10/04 -0700, [EMAIL PROTECTED] wrote: >I am looking into using sqlite for storing some data that will be ~100,000 >records in size, where each record will contain text that has an average >size of 40k, but could be > 200k. I will likely need to encrypt (and >potentially compress) the database. > > > >My question is whether to store the text in the database, or to keep a >separate file for the text with seek pointers and lengths in the database. >My preference is for the former, since I wouldn't have to manage a separate >robust encryption/compression/deletion process, but I am concerned about the >perf and size of the insertion stress experiments I've been running, and >I've seen comments on this list suggesting the latter for BLOBs. Are there >specific tuning tweaks I can make to improve my results? > > > >Thanks, > >Ken Cooper > > Sincerely, Ed Porter
