I have this query on production, which had no problem till yesterday:
SELECT E.ID
FROM EXAM_STUD_ENTERANCE E,
EXAM_REGISTRATION R
WHERE E.EXAM_REG_ID = R.ID
AND E.USER_NAME ='user'
AND E.BLOCK = 1
AND E.STATUS = 1
AND TO_DATE (TO_CHAR (SYSDATE, 'DD/MM/YYYY HH24:MI:SS'),
'DD/MM/YYYY HH24:MI:SS') >
TO_DATE (
TO_CHAR (R.EXAM_DATE, 'DD/MM/YYYY')
|| ' ' || R.EXAM_START_TIME
,
'DD/MM/YYYY HH24:MI:SS')
AND TO_DATE (TO_CHAR (SYSDATE, 'DD/MM/YYYY HH24:MI:SS'),
'DD/MM/YYYY HH24:MI:SS') <
TO_DATE (
TO_CHAR (R.EXAM_DATE, 'DD/MM/YYYY')
|| ' ' || R.EXAM_END_TIME
,
'DD/MM/YYYY HH24:MI:SS');
But now I receive an error
ORA-01858: a non-numeric character was found where a numeric was expected
Type of columns EXAM_START_TIME and EXAM_END_TIME are NVARCHAR2(100), and stores data like '16:40' ; EXAM_DATE is DATE
Yesterday someone inserted a bad value into one of those time columns
Do a
SELECT * FROM EDUMAN_EXAM.EXAM_REGISTRATION
WHERE
NOT(
REGEXP_LIKE(EXAM_START_TIME, '^\d\d:\d\d:\d\d$') AND
REGEXP_LIKE(EXAM_END_TIME, '^\d\d:\d\d:\d\d$')
)
Find the bad data, and remove it
Note: you posted an sql that indicates the time columns store hh:mm:ss and you said "it was working" but then you claimed that the time columns don't store seconds. If they do store seconds, then use the query as above. If they don't store seconds, reduce the regex pattern to \d\d:\d\d
As a side commentary, you should never have made the decision to store this time data as a string. The DATETIME that holds the EXAM_START_DATE can just as easily hold a time, and if you just want the DATE without the time part, you can TRUNC(some_datetime) to remove the time component
Store the data in a column of the proper type, and you won't get this issue, and you wont have to get into this ludicrous date -> string -> concatenate -> parse process on thousands of records