PostgreSQL: how to disable for to_tsvector reducing the tokens to lexemes

NeverSleeps 0
Viewed 15

I have such a query where I use the to_tsvector function. The documentation says:

to_tsvector parses a textual document into tokens, reduces the tokens to lexemes

But sometimes the function to_tsvector changes some names incorrectly. Is it possible for some words to cancel the cast to tokens? For example, for the value "Илья" to disable it, and for all other cases make it enable?

SELECT c.id,
       tsvector_agg(to_tsvector('russian',
                                coalesce(cv.data ->> 'name', '') || ' ' ||
                                coalesce(cv.data ->> 'surname', '') || ' ' ||
           )) as v
FROM client c
sql postgresql
1 Answers

The proper way to disable stemming for certain words is to include a synonym dictionary into the text search configuration. You'd have to add the names to the synonym file, then any word that is handled by the synonym dictionary and not processed by the stemming dictionary that comes after it.

The linked documentation gives an example for the name "Paris", but it will work just as well in your case.

Laurenz Albe0
Related