On Fri, Jun 29, 2012 at 6:09 PM, Nico Williams <[email protected]>wrote:
> On Fri, Jun 29, 2012 at 4:39 PM, Cory Nelson <[email protected]> wrote: > > On Fri, Jun 29, 2012 at 2:48 PM, Richard Hipp <[email protected]> wrote: > > What is the rationale for the 7-bit BINARY encoding? The performance > impact > > will surely outweigh any convenience of being able to treat blobs as > > 0-terminated strings. > > I tend to agree. Varint length + value should be faster, varint+value does not sort BLOBs in lexicographical order. Not having a distinct terminator for the BLOB means that two BLOBs where one is a prefix of the other might not compare correctly. > though this > just needs to be measured, and it's likely that the 7-bit approach > will be faster for very short blobs. Another possibility would be to > use 31-bit encoding ;) or escape zero-bytes, and so on. > > Nico > -- > _______________________________________________ > sqlite-users mailing list > [email protected] > http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users > -- D. Richard Hipp [email protected] _______________________________________________ sqlite-users mailing list [email protected] http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
