MySQL Bugs: #13406: DELETE FROM csv_table

Bug #13406	DELETE FROM csv_table - wrong "affected rows"
Submitted:	22 Sep 2005 15:28	Modified:	2 Dec 2005 20:32
Reporter:	Sergei Golubchik	Email Updates:
Status:	Closed	Impact on me:	None
Category:	MySQL Server	Severity:	S3 (Non-critical)
Version:		OS:
Assigned to:	Sergei Golubchik	CPU Architecture:	Any

Description:
DELETE FROM csv_table - wrong "affected rows"

How to repeat:
mysql> create table t1 (a int) type=csv;
Query OK, 0 rows affected, 1 warning (0.01 sec)

mysql> delete from t1;
Query OK, 2 rows affected (0.00 sec)

mysql> delete from t1;
Query OK, 2 rows affected (0.00 sec)

I reproduced this under FC4 / 5.0.12bk

Also, 2 is the value when there are an arbitrary number of rows of data:

mysql> insert into tinatab values (1, 1);
Query OK, 1 row affected (1.63 sec)

mysql> insert into tinatab values (2, 2);
Query OK, 1 row affected (1.64 sec)

mysql> insert into tinatab values (3,3);
Query OK, 1 row affected (2.29 sec)

mysql> delete from tinatab;
Query OK, 2 rows affected (0.89 sec)

even more issues: records is never properly known in ha_tina. It's always 0, and delete_row gives the proper number of records deleted only because it's decrementing from 0, and records is a ha_rows type which is a ulong! The reason for 2 being returned is that records is hardcoded to 2 in ::info if records < 0, which it is.

https://intranet.mysql.com/worklog/Server-Sprint/?tid=2949

Add meta file to ha_tina. This will help keep trace of the number of rows in a table, to make this bug fixable.

A patch for this bug has been committed. After review, it may
be pushed to the relevant source trees for release in the next
version. You can access the patch from:

  http://lists.mysql.com/internals/32294

A patch for this bug has been committed. After review, it may
be pushed to the relevant source trees for release in the next
version. You can access the patch from:

  http://lists.mysql.com/internals/32387

fixed in 5.0.17

Noted in 5.0.17 changelog.