Hello
I should to search distinct values based on similarity
postgres=# explain select nazobce, nazobce <-> 'Benešov' from obce
order by nazobce <-> 'Benešov' limit 10
; QUERY PLAN
-------------------------------------------------------------------------------------------Limit (cost=0.00..0.86
rows=10width=10) -> Index Scan using obce_nazobce_idx on obce (cost=0.00..1433.14
rows=16644 width=10) Order By: (nazobce <-> 'Benešov'::text)
(3 rows)
Time: 0.576 ms
but using DISTINCT breaks KNN searching optimization
postgres=# explain select distinct nazobce, nazobce <-> 'Benešov' from
obce order by nazobce <-> 'Benešov' limit 10
; QUERY PLAN
-----------------------------------------------------------------------------Limit (cost=600.45..600.47 rows=10
width=10) -> Sort (cost=600.45..613.80 rows=5341 width=10) Sort Key: ((nazobce <-> 'Benešov'::text)) ->
HashAggregate (cost=418.27..485.03 rows=5341 width=10) -> Seq Scan on obce (cost=0.00..335.05
rows=16644width=10)
(5 rows)
Regards
Pavel Stehule
