README

float4 ts_rank_tfidf(text STATTABLE [, float4[] weights] tsvector, tsquery [, int flags])

STATTABLE: Name of table stored set-wide statistic. Table should be defined as CREATE TABLE table_name ( value text UNIQUE, ndoc int4 (or bigint) NOT NULL CHECK (ndoc>0) ); And row with null value means total number of documents. See an examples in sql/setrank.sql file

example:

Create STATTABLE:

             Table "public.apod"

Column | Type | Collation | Nullable | Default ----------+----------+-----------+----------+--------- title | text | | | body | text | | | sdate | date | | | keywords | text | | | id | integer | | | fts | tsvector | | | Indexes:

select word, ndoc into apod_stat from ts_stat('select fts from apod') order by ndoc desc; insert into apod_stat values(NULL, 1754);

-- 'star' is very common word

select title, ts_rank_tfidf('apod_stat', fts, to_tsquery('x-ray:a | star:a')) as rank from apod where fts @@ to_tsquery('x-ray:a| star:a') order by rank desc limit 10; title | rank ---------------------------------------------+---------- X-ray Moon and X-ray Star | 0.218907 X-Ray Pulsar | 0.1818 The X-Ray Sky | 0.178291 X-Ray Pleiades | 0.177293 X-Ray Pleiades | 0.177293 X-Ray Pleiades | 0.177293 XMM-Newton First Light: X-Rays From The LMC | 0.176329 The X-ray Timing Explorer | 0.175857 X-Rays From The Galactic Center | 0.175857 X-Ray Jet From Centaurus A | 0.175523 (10 rows)

select title, ts_rank_cd(fts, to_tsquery('x-ray:a | star:a')) as rank from apod where fts @@ to_tsquery('x-ray:a| star:a') order by rank desc limit 10; title | rank ----------------------------+--------- X-ray Moon and X-ray Star | 3.66667 X-Ray Stars Of Orion | 2 The Pleiades Star Cluster | 1 X-Ray Jet From Centaurus A | 1 Star Wars in NGC 664 | 1 The Pleiades Star Cluster | 1 Stars in the Infrared Sky | 1 The X-Ray Moon | 1 The Pleiades Star Cluster | 1 The Crab Nebula in X-Rays | 1 (10 rows)