Reputation:
I have written the below code for splitting string, but it takes a long time to execute. Please help me re-write my code to optimize the query. I have tried hard to find the result, but i didn't get how to apply the logic to replace the cursor.
declare @table as nvarchar(50),@column as nvarchar(max),@delimiter
as nvarchar(1),@tablekey as nvarchar(max)
BEGIN
SET NOCOUNT ON;
set @table='QAT_Tsentences'
set @column='SentenceElID'
set @delimiter='|'
set @tablekey='ID'
declare @sql as nvarchar(max), @tabkey as nvarchar(max), @txt as
nvarchar(1000), @txtSplitted as nvarchar(255)
DECLARE @pos integer--, @delimiter as nchar(1)
DECLARE @Leftpiece nvarchar(255), @Rightpiece as nvarchar(255)
CREATE TABLE #t(tablekey nvarchar(max), txt nvarchar(1000))
set @sql= 'INSERT #t select '+@tablekey+','+@column+' from '+@table+' where
'+@column+' is not null'
exec(@sql)
drop table QAT_txtsplitted
CREATE TABLE QAT_txtsplitted(tablekey nvarchar(max), txt [nvarchar]
(max), txtSplitted [nvarchar](max), ID INT NOT NULL IDENTITY(1,1))
DECLARE c1 CURSOR
FOR
SELECT tablekey, txt
FROM #t
OPEN c1
FETCH NEXT FROM c1
INTO @tabkey,@txt
While (@@Fetch_status = 0)
BEGIN
SET @Rightpiece = @txt
IF RIGHT(RTRIM(@Rightpiece),1) <> @delimiter
SET @Rightpiece = @Rightpiece + @delimiter
SET @pos = CHARINDEX(@delimiter, @Rightpiece)
WHILE @pos <> 0
BEGIN
SET @Leftpiece = LEFT(@Rightpiece, @pos - 1)
INSERT INTO QAT_txtsplitted (tablekey,txt,txtsplitted) VALUES
(@tabkey,@txt, @Leftpiece);
SET @Rightpiece = STUFF(@Rightpiece, 1, @pos, '')
SET @pos = CHARINDEX(@delimiter, @Rightpiece)
END
FETCH NEXT FROM c1
INTO @tabkey,@txt
END
CLOSE c1
DEALLOCATE c1
drop table #t
nchar(1),@tablekey as decimal(15))
print 'table results : QAT_txtsplitted'
END
Please find below my result where i have tried to split for a particular tablekey.
Upvotes: 0
Views: 3031
Reputation: 993
Here is part of a SP THAT I wrote to remove a cursor
BEGIN
DECLARE @Modules TABLE
(
ID INT IDENTITY(1,1)
, ModuleId INT
, ModuleShortName NVARCHAR(256)
, ModuleTableName NVARCHAR(260)
)
INSERT INTO @Modules select ModuleId,ModuleShortName ,ModuleTableName from Mods
DECLARE @Count INT, @Counter INT, @ModuleId INT, @ModuleTableName NVARCHAR(260)
SELECT @Count = COUNT(*) FROM @Modules m
SET @Counter = 1
WHILE @Counter <= @Count
BEGIN
SELECT @ModuleId = ModuleId, @ModuleTableName = ModuleTableName FROM @Modules m WHERE ID = @Counter -- extracting values from the table by counternum
-- do something with data
-- puttin inner logic
SET @Counter = @Counter + 1
END
Upvotes: 2
Reputation: 993
Cursors can be poor performers in SQL Server which is designed for set-based operations. There are some methods to increase the performance of a cursor by applying additional arguments which can be read about here. If you only want to sequentially step through your cursor, then you could alleviate a lot of overhead by using the LOCAL FAST_FORWARD
arguements.
DECLARE C1 CURSOR LOCAL FAST_FORWARD FOR <your query>
Another method, which could perform better, is a while loop.
DECLARE @cnt INT = 0;
Declare @t_CNT = Select Count(*) from t
WHILE @cnt < @t_CNT
BEGIN
// do your work here
SET @cnt = @cnt + 1;
END;
Upvotes: 0