Description
Results in collated databases from like clauses with trailing % seem to be matching upper and lower of specified character.
For instance:
ij(CONNECTION1)> connect 'jdbc:derby:nodb;create=true;territory=no_NO;collation=
TERRITORY_BASED';
WARNING 01J01: Database 'nodb' not created, connection made to existing database
instead.
ij(CONNECTION2)> drop table t;
0 rows inserted/updated/deleted
ij(CONNECTION2)> create table t (x varchar(20));
0 rows inserted/updated/deleted
ij(CONNECTION2)> insert into t values 'Waagan', 'Wσhan', 'Wanvik', 'Wσgan', 'eks
trabetaling', 'ekstraarbeid', 'ekstra¡arbeid', '¡a', 'a', '¡b', 'b', '-a', '-b',
' a', ' b', 'A', 'B', 'C';
ij(CONNECTION2)> select * from t where x like 'a%';
X
--------------------
¡a
a
-a
a
A
5 rows selected
ij(CONNECTION2)> select * from t where x like 'b%';
X
--------------------
¡b
b
-b
b
B
5 rows selected
ij(CONNECTION2)> select * from t where x like 'c%';
X
--------------------
C
Attachments
Attachments
Issue Links
- is related to
-
DERBY-2710 disable dynamic like optimization for collated databases
- Closed