I am running the following command to delete rows in batches out of a large table (150 million rows):
DECLARE @RowCount int
WHILE 1=1
BEGIN
DELETE TOP (10000) t1
FROM table t1
INNER JOIN table2 t2 ON t2.PrimaryKey = t1.PrimaryKey
WHERE t1.YearProcessed <= 2007
SET @RowCount = @@ROWCOUNT
IF (@RowCount < 10000) BREAK
END
This table is HIGHLY used. However, it is deleting records, but it is also causing locking on some records, thus throwing errors to the user (which is not acceptable in the environment we're in).
How can I delete older records without causing locks? Should I reduce the size of the batch from 10000 records to 1000? How will this effect log sizes (we have very little hard drive space left for large log growth).
Any suggestions?
I have seen similar sporadic problems in the past where even in small batches 0f 5000 records, locking would still happen. In our case, each delete/update was contained in its own Begin Tran...Commit loop. To correct the problem, the logic of
WaitFor DELAY '00:00:00:01'
was placed at the top of each loop through and that corrected the problem.
First of all - it looks like your DELETE performing Clustered Index Scan, i recommend to do the following:
create index [IX.IndexName] ON t1(YearProcessed, PrimaryKey)
Second - is there any needs to join t2 table?
And then use following query to delete the rows, assuming that your PrimaryKey column is of type INT:
declare @ids TABLE(PrimaryKey INT)
WHILE 1=1
BEGIN
INSERT @ids
SELECT top 10000 DISTINCT t1.PrimaryKey
FROM table t1
INNER JOIN table2 t2 ON t2.PrimaryKey = t1.PrimaryKey
WHERE t1.YearProcessed <= 2007
IF @@ROWCOUNT = 0 BREAK
DELETE t1
WHERE PrimaryKey in (Select PrimaryKey from @ids)
delete from @ids
END
And do not forget to remove t2 table from join if it is not needed
If it still causes locks - then lower the amount of rows deleted in each round
I think you're on the right track.
Look at these two articles, too:
and:
http://www.dbforums.com/microsoft-sql-server/985516-deleting-without-locking.html
Before you run the delete, check the estimated query plan to see if it
is doing an index seek for the delete, or still doing a full table
scan/access.
In addition to the other suggestions (that aim at reducing the work done during deletion) you can also configure SQL Server to not block other readers while doing deletes on a table.
This can be done by using "snapshot isolation" which was introduced with SQL Server 2005:
http://msdn.microsoft.com/en-us/library/ms345124%28v=sql.90%29.aspx
If you have anything with cascading deletes make sure they are indexed.
Highlighting the DELETE query and clicking Display estimated execution plan
will show suggested indexes - which in my case included some cascading deletes.
Adding indexes for those made the delete a lot faster - but I still wouldn't try to delete all rows at once.
the best way that I have found is form asp.net DeleteExpiredSessions . you do a READUNCOMMITTED select and put the records in a temp table , than delete the record using a CURSOR.
ALTER PROCEDURE [dbo].[DeleteExpiredSessions]
AS
SET NOCOUNT ON
SET DEADLOCK_PRIORITY LOW
DECLARE @now datetime
SET @now = GETUTCDATE()
CREATE TABLE #tblExpiredSessions
(
SessionID nvarchar(88) NOT NULL PRIMARY KEY
)
INSERT #tblExpiredSessions (SessionID)
SELECT SessionID
FROM [ASPState].dbo.ASPStateTempSessions WITH (READUNCOMMITTED)
WHERE Expires < @now
IF @@ROWCOUNT <> 0
BEGIN
DECLARE ExpiredSessionCursor CURSOR LOCAL FORWARD_ONLY READ_ONLY
FOR SELECT SessionID FROM #tblExpiredSessions
DECLARE @SessionID nvarchar(88)
OPEN ExpiredSessionCursor
FETCH NEXT FROM ExpiredSessionCursor INTO @SessionID
WHILE @@FETCH_STATUS = 0
BEGIN
DELETE FROM [ASPState].dbo.ASPStateTempSessions WHERE SessionID = @SessionID AND Expires < @now
FETCH NEXT FROM ExpiredSessionCursor INTO @SessionID
END
CLOSE ExpiredSessionCursor
DEALLOCATE ExpiredSessionCursor
END
DROP TABLE #tblExpiredSessions
RETURN 0
Try this,
DECLARE @RowCount int
WHILE 1=1
BEGIN
BEGIN TRANSACTION
DELETE TOP (10000) t1
FROM table t1
INNER JOIN table2 t2 ON t2.PrimaryKey = t1.PrimaryKey
WHERE t1.YearProcessed <= 2007
END TRANSACTION
COMMIT TRANSACTION
SET @RowCount = @@ROWCOUNT
IF (@RowCount < 10000) BREAK
END