I have a table of the structure:
+---------+--------------+-----------------+---------------+-------+------+ | week_no | long_week_no | week_start_date | week_end_date | month | year | +---------+--------------+-----------------+---------------+-------+------+ | 1 | 1A | 01/01/2015 | 03/01/2015 | JAN | 2015 | | 1 | 1B | 04/01/2015 | 10/01/2015 | JAN | 2015 | | 2 | 2 | 11/01/2015 | 17/01/2015 | JAN | 2015 | | 3 | 3 | 18/01/2015 | 24/01/2015 | JAN | 2015 | | .. | .. | .. | .. | .. | .. | | 51 | 51 | 14/12/2014 | 20/12/2015 | DEC | 2014 | +---------+--------------+-----------------+---------------+-------+------+
When I run the following statement:
SELECT *
FROM loy_period
WHERE To_date('15/04/2015', 'DD/MM/YYYY') BETWEEN
To_date(week_start_date, 'DD/MM/YYYY') AND
To_date(week_end_date, 'DD/MM/YYYY');
and
SELECT *
FROM loy_period
WHERE To_date('15/04/2015', 'DD/MM/YYYY') BETWEEN
week_start_date AND
week_end_date;
It returns the following:
+---------+--------------+-----------------+---------------+-------+------+ | week_no | long_week_no | week_start_date | week_end_date | month | year | +---------+--------------+-----------------+---------------+-------+------+ | 51 | 51 | 14/12/2014 | 20/12/2015 | DEC | 2014 | | 1 | 1A | 01/01/2015 | 03/01/2015 | JAN | 2015 | +---------+--------------+-----------------+---------------+-------+------+
And when I run it with any date, it will return the correct period in addition to the week_no 51 record!
I'm confused why this is happening. Both columns week_start_date
and week_end_date
of the type date
.
Fiddle works as expected.
It works perfectly for me with proper use of TO_DATE and DATE values.
Never use TO_DATE on a DATE, It will implicitly convert it into string and then back to date using locale-specific NLS format.
'01/01/2015'
is NOT a DATE, it is a STRING. You must use TO_DATE to explicitly convert it into DATE.
See what happens:
SQL> explain plan for select * from dual where to_date(sysdate) > to_date(sysdate -1);
Explained.
SQL> select * from table(dbms_xplan.display);
PLAN_TABLE_OUTPUT
--------------------------------------------------------------------------------
Plan hash value: 3752461848
---------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
---------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 1 | 2 | 2 (0)| 00:00:01 |
|* 1 | FILTER | | | | | |
| 2 | TABLE ACCESS FULL| DUAL | 1 | 2 | 2 (0)| 00:00:01 |
---------------------------------------------------------------------------
Predicate Information (identified by operation id):
PLAN_TABLE_OUTPUT
--------------------------------------------------------------------------------
---------------------------------------------------
1 - filter(TO_DATE(TO_CHAR(SYSDATE@!))>TO_DATE(TO_CHAR(SYSDATE@!-1)))
14 rows selected.
SQL>
So, the actual filter applied is filter(TO_DATE(TO_CHAR(SYSDATE@!)) You will get incorrect output due to the implicit conversion based on locale-specific NLS format.
Anyway, now coming back to your original question.
For example,
Let's say your data looks like:
Setup:
SQL> CREATE TABLE t
2 (week_no VARCHAR2(2), long_week_no VARCHAR2(2), week_start_date DATE, week_end_date DATE, mnth VARCHAR2(3), yr VARCHAR2(4))
3 ;
Table created.
SQL>
SQL>
SQL> INSERT ALL
2 INTO t (week_no, long_week_no, week_start_date, week_end_date, mnth, yr)
3 VALUES ('1', '1A', TO_DATE('01/01/2015','DD/MM/YYYY'), TO_DATE('03/01/2015','DD/MM/YYYY'), 'JAN', '2015')
4 INTO t (week_no, long_week_no, week_start_date, week_end_date, mnth, yr)
5 VALUES ('1', '1B', TO_DATE('04/01/2015','DD/MM/YYYY'), TO_DATE('10/01/2015','DD/MM/YYYY'), 'JAN', '2015')
6 INTO t (week_no, long_week_no, week_start_date, week_end_date, mnth, yr)
7 VALUES ('2', '2', TO_DATE('11/01/2015','DD/MM/YYYY'), TO_DATE('17/01/2015','DD/MM/YYYY'), 'JAN', '2015')
8 INTO t (week_no, long_week_no, week_start_date, week_end_date, mnth, yr)
9 VALUES ('3', '3', TO_DATE('18/01/2015','DD/MM/YYYY'), TO_DATE('24/01/2015','DD/MM/YYYY'), 'JAN', '2015')
10 INTO t (week_no, long_week_no, week_start_date, week_end_date, mnth, yr)
11 VALUES ('51', '51', TO_DATE('20/12/2014','DD/MM/YYYY'), TO_DATE('26/12/2015','DD/MM/YYYY'), 'DEC', '2014')
12 SELECT * FROM dual
13 ;
5 rows created.
SQL>
SQL> COMMIT;
Commit complete.
SQL>
Table:
SQL> SELECT * FROM t;
WE LO WEEK_STAR WEEK_END_ MNT YR
-- -- --------- --------- --- ----
1 1A 01-JAN-15 03-JAN-15 JAN 2015
1 1B 04-JAN-15 10-JAN-15 JAN 2015
2 2 11-JAN-15 17-JAN-15 JAN 2015
3 3 18-JAN-15 24-JAN-15 JAN 2015
51 51 20-DEC-14 26-DEC-15 DEC 2014
SQL>
Query to filter the rows based on DATE range:
SQL> SELECT *
2 FROM t
3 WHERE To_date('15/01/2015', 'DD/MM/YYYY') BETWEEN
4 week_start_date AND
5 week_end_date;
WE LO WEEK_STAR WEEK_END_ MNT YR
-- -- --------- --------- --- ----
2 2 11-JAN-15 17-JAN-15 JAN 2015
51 51 20-DEC-14 26-DEC-15 DEC 2014
SQL>
来源:https://stackoverflow.com/questions/29558899/date-comparison-returns-unusual-result-sql-oracle