group by

Dear Developers,

There is a possible bug in 'select ... group by' SQL.
I reported it on the bugreport form on the web. (I think it number is
2416?). But no any reaction to it.
It is not problem for me, I make my results in other way. But it is
possible problem for other pepoples.

I reproduced it in smaller database table in other way. (with 180000
records) (postgresql 8.03 on linux)
e.g.:
there is a table:
stat-# \d summary
Table "public.summary"
colum | Type | MĂłdosĂ­tĂł
-----------+------------------------+----------
kifejezes | character varying(300) | not null
cnt | integer | not null
talalat | integer |
Indexes:
"idx_summary_cnt" btree (cnt) CLUSTER
"idx_summary_kifejezes" btree (kifejezes text_pattern_ops)

1)
select kifejezes, count(kifejezes) from summary group by kifejezes
having count(kifejezes)>1;
the result is:
kifejezes | count
-------------------------+-------
csĂşcscsajok | 2
jĂĄszszentandrĂĄs | 3
kullancscsĂ­pĂŠs | 2
magannyugdijpenztar | 2
magĂĄnnyugdijpĂŠnztĂĄr | 2
magĂĄnnyugdĂ­jpĂŠnztĂĄr | 3
magĂĄnnyugdĂ­jpĂŠnztĂĄrak | 2
mĹąvelĹdĂŠsszociolĂłgia | 2
otp magĂĄnnyugdĂ­jpĂŠnztĂĄr | 2
(9 rows)

2)
select * from summary where kifejezes like 'jegygy%';
kifejezes | cnt | talalat
------------+-----+---------
jegygyĂźrĹą | 4 | 0
jegygyĹąrĹą | 5 | 0
jegygyĹąrĹą | 7 | 0
jegygyĹąrĹą | 12 | 0
jegygyĹąrĹąk | 3 | 0
(5 rows)

Why not is in the first query results the "jegygyĹąrĹą" (second query
rows )?