*** 2,7 **** --- 2,13 ---- ---- + *Integrate fts2 storage with sqlite btree* + + Convert %_segments to have fields for level, idx, and term0, and pull term0 out of the leaf nodes. In this form, interior nodes dissappear entirely, as does the segment directory. Segment merges would be pretty much unchanged. This could be a substantial improvement when implementing incremental merging, since it removes a layer of management from the system. Main concern is whether the additional btree will slow things down too much - the test immediately before fts2 used parallel btrees, and was still 10x slower than fts2 ended up. + + ---- + *Base docids per segment* Delta-encoding of docids kicks in strongly for longer doclists, but doesn't help much for short doclists. In the limit, a single-document segment for a high docid will spend much more overhead to encode identical docids. A solution would be to have a per-segment base docid which all doclists in that segment delta-encode from. For a single-document segment, this would make all leading docids 0, which could then be omitted.
Small. Fast. Reliable.
Choose any three.
Choose any three.
|
|