[Gnumed-devel] darn ! sensitive vs. insensitive searching

[Karsten Hilbert]

Can you please try running this against your version of
PostgreSQL ? I am using 7.1:

 gm-archive=# explain select id, lastnames, firstnames from names where 
lastnames ~ '^becker' and firstnames ~ '^k';
  NOTICE:  QUERY PLAN:
  Index Scan using idx_names_last_first on names
  (cost=0.00..2.54 rows=1 width=28)
  EXPLAIN

 gm-archive=# explain select id, lastnames, firstnames from names where 
lastnames ~* '^becker' and firstnames ~* '^k';
  NOTICE:  QUERY PLAN:
  Seq Scan on names  (cost=0.00..1821.34 rows=1 width=28)
  EXPLAIN

 gm-archive=#

How can we work around this ? Is it reasonable to assume that
the user always types in the *beginning* of the name parts ?

IOW, to the users amongst us: Do we almost always type in the
starting letters of a name when we search for it ? Or do we
often search for arbitrary parts inside a name ? I'd assume
not so.

Karsten
-- 
GPG key ID E4071346 @ wwwkeys.pgp.net
E167 67FD A291 2BEA 73BD  4537 78B9 A9F9 E407 1346