I have the following table:
so2, which has following column details:
id, title, posts, body (tsvector).
And I created the index on the following:
"so2_pkey" PRIMARY KEY, btree (id)
"body" gin (body)
And I wanted to query on my tsvector with the string: `Is it possible to toggle the visibility of an element`. I tried the following query:
SELECT id, ts_rank(body, keywords) AS rank
FROM so2,
to_tsquery('Is & it & possible & to & toggle & the & visibility & of & an & element')
keywords
WHERE keywords @@ body ORDER BY rank DESC;
This give the result in which the "element" string count is large in number, however there is a record which exactly matches the search string, that comes as third in my result. So logically it just checked in my query which word count occurs the most across my body column, which I don't want.
Can I say somehow to postgresql, that I want the pharse search rather than the text spilt up weight in the document. Is it possible to do?
Edit: I tried with `ts_rank_cd\` as well. Same result.