> We need to produce copies of our databases for archive. > It is a requirement that the size of those copies being as small as > possible, without having to perform an external compression. > vacuum doesn't seem to perform a compression (it works on fragmented > data), is there any other way to do that ?
If you can't use an external compression program (which would almost certainly help reduce the size of your archived database), then there are a couple of options I can think of: 1. When you create the copy of your database, you could drop all of the indices from the copy, then vacuum. Depending on your schema, this has the potential to remove some redundant information. (At the expense of query speed, of course.) You could always re-create the indices, if needed, when reading the archive. 2. If that doesn't help enough, run the sqlite3_analyzer (from http://sqlite.org/download.html) to see which table(s) are using the most disk space. Focus on these tables to see if you can save space: Can you better normalize the schema to reduce repeated values? Can some (non-vital) data be omitted from the archive? etc. If the above two options don't help enough, than I would reconsider the external compression tool. zlib, for example, is a relatively lightweight, open source compression library that may do well on your database. Hope this helps, Eric _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
