Re: Fuzzy substring searching with the pg_trgm extension

2016-03-18 23:46 GMT+03:00 Jeff Janes <jeff(dot)janes(at)gmail(dot)com>:
>
>
> <% and <<-> are not documented at all. Is that a deliberate choice?
> Since they were added as convenience functions for the user, I think
> they really need to be in the user documentation.
>

I can send a patch a little bit later. I documented %>
and <->> because examples of other operators have the following order:

SELECT t, t <-> 'word' AS dist
FROM test_trgm
ORDER BY dist LIMIT 10;

and

SELECT * FROM test_trgm WHERE t LIKE '%foo%bar';

I did not include <% and <<-> because I did not know how to document
commutators. But I can fix it.

And honestly the following order:

SELECT 'word' <% t FROM test_trgm;

is more convenient to me too.

Do you know how do not break the line in the operators table in the first
column? Now I see:

Operator | Returns
----------------|------------------
text % | boolean
text |

But the following is better:

Operator | Returns
----------------|------------------
text % text | boolean

>
> Also, the documentation should probably include <% and <<-> as the
> "parent" operators and use them in the examples, and only mention %>
> and <->> in passing, as the commutators. That is because <% and <<->
> take their arguments in the same order as word_similarity does. It
> would be less confusing if the documentation and the examples didn't
> need to keep changing the argument orders.
>
> Cheers,
>
> Jeff
>

--
Artur Zakirov
Postgres Professional: http://www.postgrespro.com
Russian Postgres Company