Compression in the DB is interesting I think the commercial prod mentioned just does a field compress and that is all. In general this only works on larger blob like fiels as the overhead of the compressor is usually somewhat high and lets not forget extra overhead of comp/decomp. The idea I was playing with a while back (zlib) was a global db dictionary for compression, but as memory got cheep and larger I dropped the project. The simple token compression (like the old days of faircom's btree package, sybase IQ, Monet, etc are much nicer). It would be a neat feature.
Sandy >> My question is (and I've raised this topic back in September, but >> didn't get back to it since), does anyone have a free/commercial >> add-on for SQLite v3 to perform on-the-fly compression/decompression >> of data, preferably on a field level (compress just one of the fields, >> not the whole table)? > > Related to this I would love to see reference counting of values. > For example if I add the string "foobar" in 27 different places, > it only gets stored once with a reference count of 27. > > There are various places that have done compression: > > http://www.sqliteplus.com/ > > There is also mention of compression at > > http://www.hwaci.com/sw/sqlite/prosupport.html > > If you are working on a commercial product and SQLite has made your > product better and/or improved your development process then it is > fair and worthwhile to pay for that. > > Roger >
