Teodor Sigaev kirjutas K, 10.12.2003 kell 11:20: > >>Tsearch was never minded as prefix search, and index structure doesn't support > >>any kind of prefix or suffix. But you can write extension to tsearch, which will > >>search by prefix. But such solution wiil not use index, only sequence scan. > > > > > > How efficient would tsearch be for really big expressions (where 'hu%' > > would be expanded (using a btree word index on one column word table) to > > tsearch equivalent of ( "human" or "humanity" or "humming" or "huge" or > > ..1000 words here...) before passing the expression to tsearch? > > GiST index of tsearch doen't support prefix search, so it will works only by > seqscan, as we know :) disk is much more slow than processor, speed will be > limited by disk.
I meant that the expansion of 'hu%' is done before and outside of tsearch, so the question is how efficient will tsearch be for searching for hudreds or thousands of words in one expression. > > How hard (or sensible ;) would be creating such an index using GiST ? > > As proved by tsearch GiST can cope well with many-to-many indexes. > > Sorry, I don't understand. Do you mean that GiST supports one heap tuple in > several index tuple? If yes then no :). GiST doesn't support this feature. I > don't think that GiST may help in this situation. but tsearch seems to support this, and tsearch uses GiST. Is this functionality added entirely by tsearch ? -------------- Hannu ---------------------------(end of broadcast)--------------------------- TIP 7: don't forget to increase your free space map settings
