I have the following table:
CREATE TABLE test (key1 varchar(10), key2 varchar(10), col3 varchar(10),
col4 varchar(10), col5 varchar(10), val varchar(10))
INSERT INTO test VALUES('A', 'B', 'C', 'D', 'F', 'good1')
INSERT INTO test VALUES('F', 'C', 'C', 'D', 'F', 'bad1')
INSERT INTO test VALUES('A', 'D', 'C', 'D', 'F', 'good2')
INSERT INTO test VALUES('N', 'B', 'C', 'D', 'F', 'bad2')
INSERT INTO test VALUES('A', 'B', 'C', 'D', 'F', 'want this')
INSERT INTO test VALUES('A', 'D', 'C', 'D', 'F', 'and this')
I would like to extract key1, key2 pairs that where val column includes "good". Then I need to go back to the original table and extract all the entries where the mentioned key1, key2 pairs appear.
I am using the following:
select t.* from test t where exists (select 1
from t t2
where t2.value like '%good%') and
t2.key1 = t.key1 and
t2.key2 = t.key2)
Upon running the above, I get the following error:
"Object t does not exist"