On 13/5/2014 2:45 PM, Simon Slavin wrote:
There are two versions of your question: one for compression of a database which is only going to be read from now on and another for compression of a database which has to support writing. hwaci, Richard Hipp's own company, support both ZIPVFS and CEROD: http://www.hwaci.com/sw/sqlite/zipvfs.html http://www.hwaci.com/sw/sqlite/cerod.html which do both things. Since the guy who runs that company is also the main man behind SQLite itself, there's a good chance that support will continue. When comparing database sizes and compression factors you should bear in mind that different DBMSs access data at different speeds and require more or fewer indexes to do it at the same speeds. In other words, publishing a simpler '40% of file size' doesn't tell the whole story.
While Simon is absolutely correct, I can confirm that in production, our databases using CEROD (read only) are typically around 40% ~ 45% of the size of the original database. However, we have also seen cases, where we reduce the size of the original database by 200MB and the resulting compressed database size reduces only by 20MB since the original data had a lot of data that was very easy to compress.
As always, your mileage may vary - but this is what we have observed with CEROD.
Best Regards, Mohit. _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
