I have table in which the data is been continuously added at a rapid pace.
And i need to fetch record from this table and immediately remove them so i cannot process the same record second time. And since the data is been added at a faster rate, i need to use the TOP clause so only small number of records go to business logic for processing at the time.
I am using the below query to
BEGIN TRAN readrowdata
SELECT
top 5 [RawDataId],
[RawData]
FROM
[TABLE] with(HOLDLOCK)
WITH q AS
(
SELECT
top 5 [RawDataId],
[RawData]
FROM
[TABLE] with(HOLDLOCK)
)
DELETE from q
COMMIT TRANSACTION readrowdata
I am using the HOLDLOCK here, so new data cannot insert into the table while i am performing the SELECT and DELETE operation. I used it because Suppose if there are only 3 records in the table now, so the SELECT statement will get 3 records and in the same time new record gets inserted and the DELETE statement will delete 4 records. So i will loose 1 data here.
Is the query is ok in performance term? If i can improve it then please provide me your suggestion.
Thank you
Personally, I'd use a different approach. One with less locking, but also extra information signifying that certain records are currently being processed...
DECLARE #rowsBeingProcessed TABLE (
id INT
);
WITH rows AS (
SELECT top 5 [RawDataId] FROM yourTable WHERE processing_start IS NULL
)
UPDATE rows SET processing_start = getDate() WHERE processing_start IS NULL
OUTPUT INSERTED.RowDataID INTO #rowsBeingProcessed;
-- Business Logic Here
DELETE yourTable WHERE RowDataID IN (SELECT id FROM #rowsBeingProcessed);
Then you can also add checks like "if a record has been 'beingProcessed' for more than 10 minutes, assume that the business logic failed", etc, etc.
By locking the table in this way, you force other processes to wait for your transaction to complete. This can have very rapid consequences on scalability and performance - and it tends to be hard to predict, because there's often a chain of components all relying on your database.
If you have multiple clients each running this query, and multiple clients adding new rows to the table, the overall system performance is likely to deteriorate at some times, as each "read" client is waiting for a lock, the number of "write" clients waiting to insert data grows, and they in turn may tie up other components (whatever is generating the data you want to insert).
Diego's answer is on the money - put the data into a variable, and delete matching rows. Don't use locks in SQL Server if you can possibly avoid it!
You can do it very easily with TRIGGERS. Below mentioned is a kind of situation which will help you need not to hold other users which are trying to insert data simultaneously. Like below...
Data Definition language
CREATE TABLE SampleTable
(
id int
)
Sample Record
insert into SampleTable(id)Values(1)
Sample Trigger
CREATE TRIGGER SampleTableTrigger
on SampleTable AFTER INSERT
AS
IF Exists(SELECT id FROM INSERTED)
BEGIN
Set NOCOUNT ON
SET XACT_ABORT ON
Begin Try
Begin Tran
Select ID From Inserted
DELETE From yourTable WHERE ID IN (SELECT id FROM Inserted);
Commit Tran
End Try
Begin Catch
Rollback Tran
End Catch
End
Hope this is very simple and helpful
If I understand you correctly, you are worried that between your select and your delete, more records would be inserted and the first TOP 5 would be different then the second TOP 5?
If that so, why don't you load your first select into a temp table or variable (or at least the PKs) do whatever you have to do with your data and then do your delete based on this table?
I know that it's old question, but I found some solution here https://www.simple-talk.com/sql/learn-sql-server/the-delete-statement-in-sql-server/:
DECLARE #Output table
(
StaffID INT,
FirstName NVARCHAR(50),
LastName NVARCHAR(50),
CountryRegion NVARCHAR(50)
);
DELETE SalesStaff
OUTPUT DELETED.* INTO #Output
FROM Sales.vSalesPerson sp
INNER JOIN dbo.SalesStaff ss
ON sp.BusinessEntityID = ss.StaffID
WHERE sp.SalesLastYear = 0;
SELECT * FROM #output;
Maybe it will be helpfull for you.
Related
For a sync process, my SQL Server database should record a list items that have changed - table name and primary key.
The DB already has a table and stored procedure to do this:
EXEC #ErrCode = dbo.SyncQueueItem "tableName", 1234;
I'd like to add triggers to a table to call this stored procedure on INSERT, UPDATE, DELETE. How do I get the key? What's the simplest thing that could possibly work?
CREATE TABLE new_employees
(
id_num INT IDENTITY(1,1),
fname VARCHAR(20),
minit CHAR(1),
lname VARCHAR(30)
);
GO
IF OBJECT_ID ('dbo.sync_new_employees','TR') IS NOT NULL
DROP TRIGGER sync_new_employees;
GO
CREATE TRIGGER sync_new_employees
ON new_employees
AFTER INSERT, UPDATE, DELETE
AS
DECLARE #Key Int;
DECLARE #ErrCode Int;
-- How to get the key???
SELECT #Key = 12345;
EXEC #ErrCode = dbo.SyncQueueItem "new_employees", #key;
GO
The way to access the records changed by the operation is by using the Inserted and Deleted pseudo-tables that are provided to you by SQL Server.
Inserted contains any inserted records, or any updated records with their new values.
Deleted contains any deleted records, or any updated records with their old values.
More Info
When writing a trigger, to be safe, one should always code for the case when multiple records are acted upon. Unfortunately if you need to call a SP that means a loop - which isn't ideal.
The following code shows how this could be done for your example, and includes a method of detecting whether the operation is an Insert/Update/Delete.
declare #Key int, #ErrCode int, #Action varchar(6);
declare #Keys table (id int, [Action] varchar(6));
insert into #Keys (id, [Action])
select coalesce(I.id, D.id_num)
, case when I.id is not null and D.id is not null then 'Update' when I.id is not null then 'Insert' else 'Delete' end
from Inserted I
full join Deleted D on I.id_num = D.id_num;
while exists (select 1 from #Keys) begin
select top 1 #Key = id, #Action = [Action] from #Keys;
exec #ErrCode = dbo.SyncQueueItem 'new_employees', #key;
delete from #Keys where id = #Key;
end
Further: In addition to solving your specified problem its worth noting a couple of points regarding the bigger picture.
As #Damien_The_Unbeliever points out there are built in mechanisms to accomplish change tracking which will perform much better.
If you still wish to handle your own change tracking, it would perform better if you could arrange it such that you handle the entire recordset in one go as opposed to carrying out a row-by-row operation. There are 2 ways to accomplish this a) Move your change tracking code inside the trigger and don't use a SP. b) Use a "User Defined Table Type" to pass the record-set of changes to the SP.
You should use the Magic Table to get the data.
Usually, inserted and deleted tables are called Magic Tables in the context of a trigger. There are Inserted and Deleted magic tables in SQL Server. These tables are automatically created and managed by SQL Server internally to hold recently inserted, deleted and updated values during DML operations (Insert, Update and Delete) on a database table.
Inserted magic table
The Inserted table holds the recently inserted values, in other words, new data values. Hence recently added records are inserted into the Inserted table.
Deleted magic table
The Deleted table holds the recently deleted or updated values, in other words, old data values. Hence the old updated and deleted records are inserted into the Deleted table.
**You can use the inserted and deleted magic table to get the value of id_num **
SELECT top 1 #Key = id_num from inserted
Note: This code sample will only work for a single record for insert scenario. For Bulk insert/update scenarios you need to fetch records from inserted and deleted table stored in the temp table or variable and then loop through it to pass to your procedure or you can pass a table variable to your procedure and handle the multiple records there.
A DML trigger should operate set data else only one row will be processed. It can be something like this. And of course use magic tables inserted and deleted.
CREATE TRIGGER dbo.tr_employees
ON dbo.employees --the table from Northwind database
AFTER INSERT,DELETE,UPDATE
AS
BEGIN
-- SET NOCOUNT ON added to prevent extra result sets from
-- interfering with SELECT statements.
SET NOCOUNT ON;
declare #tbl table (id int identity(1,1),delId int,insId int)
--Use "magic tables" inserted and deleted
insert #tbl(delId, insId)
select d.EmployeeID, i.EmployeeID
from inserted i --empty when "delete"
full join deleted d --empty when "insert"
on i.EmployeeID=d.EmployeeID
declare #id int,#key int,#action char
select top 1 #id=id, #key=isnull(delId, insId),
#action=case
when delId is null then 'I'
when insId is null then 'D'
else 'U' end --just in case you need the operation executed
from #tbl
--do something for each row
while #id is not null --instead of cursor
begin
--do the main action
--exec dbo.sync 'employees', #key, #action
--remove processed row
delete #tbl where id=#id
--refill #variables
select top 1 #id=id, #key=isnull(delId, insId),
#action=case
when delId is null then 'I'
when insId is null then 'D'
else 'U' end --just in case you need the operation executed
from #tbl
end
END
Not the best solution, but just a direct answer on the question:
SELECT #Key = COALESCE(deleted.id_num,inserted.id_num);
Also not the best way (if not the worst) (do not try this at home), but at least it will help with multiple values:
DECLARE #Key INT;
DECLARE triggerCursor CURSOR LOCAL FAST_FORWARD READ_ONLY
FOR SELECT COALESCE(i.id_num,d.id_num) AS [id_num]
FROM inserted i
FULL JOIN deleted d ON d.id_num = i.id_num
WHERE (
COALESCE(i.fname,'')<>COALESCE(d.fname,'')
OR COALESCE(i.minit,'')<>COALESCE(d.minit,'')
OR COALESCE(i.lname,'')<>COALESCE(d.lname,'')
)
;
OPEN triggerCursor;
FETCH NEXT FROM triggerCursor INTO #Key;
WHILE ##FETCH_STATUS = 0
BEGIN
EXEC #ErrCode = dbo.SyncQueueItem 'new_employees', #key;
FETCH NEXT FROM triggerCursor INTO #Key;
END
CLOSE triggerCursor;
DEALLOCATE triggerCursor;
Better way to use trigger based "value-change-tracker":
INSERT INTO [YourTableHistoryName] (id_num, fname, minit, lname, WhenHappened)
SELECT COALESCE(i.id_num,d.id_num) AS [id_num]
,i.fname,i.minit,i.lname,CURRENT_TIMESTAMP AS [WhenHeppened]
FROM inserted i
FULL JOIN deleted d ON d.id_num = i.id_num
WHERE ( COALESCE(i.fname,'')<>COALESCE(d.fname,'')
OR COALESCE(i.minit,'')<>COALESCE(d.minit,'')
OR COALESCE(i.lname,'')<>COALESCE(d.lname,'')
)
;
The best (in my opinion) way to track changes is to use Temporal tables (SQL Server 2016+)
inserted/deleted in triggers will generate as many rows as touched and calling a stored proc per key would require a cursor or similar approach per row.
You should check timestamp/rowversion in SQL Server. You could add that to the all tables in question (not null, auto increment, unique within database for each table/row etc).
You could add a unique index on that column to all tables you added the column.
##DBTS is the current timestamp, you can store today's ##DBTS and tomorrow you will scan all tables from that to current ##DBTS. timestamp/rowversion will be incremented for all updates and inserts but for deletes it won't track, for deletes you can have a delete only trigger and insert keys into a different table.
Change data capture or change tracking could do this easier, but if there is heavy volumes on the server or large number of data loads, partition switches scanning the transaction log becomes a bottleneck and in some cases you will have to remove change data capture to save the transaction log from growing indefinetely.
I'm developing a site in asp.net MVC that should generate invoices.
After some reading regarding the generation of the invoice numbers, my understading is to use a trigger or a stored procedure to be sure to have the number generated correctly without skipped or duplicated numbers, due to concurrent inserts of invoices.
For what I understand, the best approach would be to create a trigger, in the after insert of my invoices table, that does the work in a single transaction.
So I came up with this, that seems to work nice (more tests this weekend)
CREATE TRIGGER Invoices_SetInvoiceNum ON Invoices AFTER INSERT
AS
declare #next int
Begin tran
set nocount on
update counters set #next=InvoiceNum=InvoiceNum+1
update Invoices set InvoiceNum = #next from inserted i join Invoices p on p.Id = i.Id
IF ##ERROR = 0
COMMIT TRAN
ELSE
ROLLBACK TRAN
Is this a good approach or there is a better way to be sure that the number will never be duplicated or skipped?
It's more safe to use also a table/record lock? if yes could you suggest an integration to my trigger?
Thanks in advance for any insight on this topic.
A better approach would be make the invoice column a Identity column and let it increment itself.
CREATE TABLE Invoices
(
InvoiceNum int IDENTITY(1,1),
AccountNum int,
OtherStuff varchar(30)
);
See this runnable example
I have a stored procedure that is responsible for inserting or updating multiple records at once. I want to perform this in my stored procedure for the sake of performance.
This stored procedure takes in a comma-delimited list of permit IDs and a status. The permit IDs are stored in a variable called #PermitIDs. The status is stored in a variable called #Status. I have a user-defined function that converts this comma-delimited list of permit IDs into a Table. I need to go through each of these IDs and do either an insert or update into a table called PermitStatus.
If a record with the permit ID does not exist, I want to add a record. If it does exist, I'm want to update the record with the given #Status value. I know how to do this for a single ID, but I do not know how to do it for multiple IDs. For single IDs, I do the following:
-- Determine whether to add or edit the PermitStatus
DECLARE #count int
SET #count = (SELECT Count(ID) FROM PermitStatus WHERE [PermitID]=#PermitID)
-- If no records were found, insert the record, otherwise add
IF #count = 0
BEGIN
INSERT INTO
PermitStatus
(
[PermitID],
[UpdatedOn],
[Status]
)
VALUES
(
#PermitID,
GETUTCDATE(),
1
)
END
ELSE
UPDATE
PermitStatus
SET
[UpdatedOn]=GETUTCDATE(),
[Status]=#Status
WHERE
[PermitID]=#PermitID
How do I loop through the records in the Table returned by my user-defined function to dynamically insert or update the records as needed?
create a split function, and use it like:
SELECT
*
FROM YourTable y
INNER JOIN dbo.splitFunction(#Parameter) s ON y.ID=s.Value
I prefer the number table approach
For this method to work, you need to do this one time table setup:
SELECT TOP 10000 IDENTITY(int,1,1) AS Number
INTO Numbers
FROM sys.objects s1
CROSS JOIN sys.objects s2
ALTER TABLE Numbers ADD CONSTRAINT PK_Numbers PRIMARY KEY CLUSTERED (Number)
Once the Numbers table is set up, create this function:
CREATE FUNCTION [dbo].[FN_ListToTableAll]
(
#SplitOn char(1) --REQUIRED, the character to split the #List string on
,#List varchar(8000)--REQUIRED, the list to split apart
)
RETURNS TABLE
AS
RETURN
(
----------------
--SINGLE QUERY-- --this WILL return empty rows
----------------
SELECT
ROW_NUMBER() OVER(ORDER BY number) AS RowNumber
,LTRIM(RTRIM(SUBSTRING(ListValue, number+1, CHARINDEX(#SplitOn, ListValue, number+1)-number - 1))) AS ListValue
FROM (
SELECT #SplitOn + #List + #SplitOn AS ListValue
) AS InnerQuery
INNER JOIN Numbers n ON n.Number < LEN(InnerQuery.ListValue)
WHERE SUBSTRING(ListValue, number, 1) = #SplitOn
);
GO
You can now easily split a CSV string into a table and join on it:
select * from dbo.FN_ListToTableAll(',','1,2,3,,,4,5,6777,,,')
OUTPUT:
RowNumber ListValue
----------- ----------
1 1
2 2
3 3
4
5
6 4
7 5
8 6777
9
10
11
(11 row(s) affected)
To make what you need work, do the following:
--this would be the existing table
DECLARE #OldData table (RowID int, RowStatus char(1))
INSERT INTO #OldData VALUES (10,'z')
INSERT INTO #OldData VALUES (20,'z')
INSERT INTO #OldData VALUES (30,'z')
INSERT INTO #OldData VALUES (70,'z')
INSERT INTO #OldData VALUES (80,'z')
INSERT INTO #OldData VALUES (90,'z')
--these would be the stored procedure input parameters
DECLARE #IDList varchar(500)
,#StatusList varchar(500)
SELECT #IDList='10,20,30,40,50,60'
,#StatusList='A,B,C,D,E,F'
--stored procedure local variable
DECLARE #InputList table (RowID int, RowStatus char(1))
--convert input prameters into a table
INSERT INTO #InputList
(RowID,RowStatus)
SELECT
i.ListValue,s.ListValue
FROM dbo.FN_ListToTableAll(',',#IDList) i
INNER JOIN dbo.FN_ListToTableAll(',',#StatusList) s ON i.RowNumber=s.RowNumber
--update all old existing rows
UPDATE o
SET RowStatus=i.RowStatus
FROM #OldData o WITH (UPDLOCK, HOLDLOCK) --to avoid race condition when there is high concurrency as per #emtucifor
INNER JOIN #InputList i ON o.RowID=i.RowID
--insert only the new rows
INSERT INTO #OldData
(RowID, RowStatus)
SELECT
i.RowID, i.RowStatus
FROM #InputList i
LEFT OUTER JOIN #OldData o ON i.RowID=o.RowID
WHERE o.RowID IS NULL
--display the old table
SELECT * FROM #OldData order BY RowID
OUTPUT:
RowID RowStatus
----------- ---------
10 A
20 B
30 C
40 D
50 E
60 F
70 z
80 z
90 z
(9 row(s) affected)
EDIT thanks to #Emtucifor click here for the tip about the race condition, I have included the locking hints in my answer, to prevent race condition problems when there is high concurrency.
There are various methods to accomplish the parts you ask are asking about.
Passing Values
There are dozens of ways to do this. Here are a few ideas to get you started:
Pass in a string of identifiers and parse it into a table, then join.
SQL 2008: Join to a table-valued parameter
Expect data to exist in a predefined temp table and join to it
Use a session-keyed permanent table
Put the code in a trigger and join to the INSERTED and DELETED tables in it.
Erland Sommarskog provides a wonderful comprehensive discussion of lists in sql server. In my opinion, the table-valued parameter in SQL 2008 is the most elegant solution for this.
Upsert/Merge
Perform a separate UPDATE and INSERT (two queries, one for each set, not row-by-row).
SQL 2008: MERGE.
An Important Gotcha
However, one thing that no one else has mentioned is that almost all upsert code, including SQL 2008 MERGE, suffers from race condition problems when there is high concurrency. Unless you use HOLDLOCK and other locking hints depending on what's being done, you will eventually run into conflicts. So you either need to lock, or respond to errors appropriately (some systems with huge transactions per second have used the error-response method successfully, instead of using locks).
One thing to realize is that different combinations of lock hints implicitly change the transaction isolation level, which affects what type of locks are acquired. This changes everything: which other locks are granted (such as a simple read), the timing of when a lock is escalated to update from update intent, and so on.
I strongly encourage you to read more detail on these race condition problems. You need to get this right.
Conditional Insert/Update Race Condition
“UPSERT” Race Condition With MERGE
Example Code
CREATE PROCEDURE dbo.PermitStatusUpdate
#PermitIDs varchar(8000), -- or (max)
#Status int
AS
SET NOCOUNT, XACT_ABORT ON -- see note below
BEGIN TRAN
DECLARE #Permits TABLE (
PermitID int NOT NULL PRIMARY KEY CLUSTERED
)
INSERT #Permits
SELECT Value FROM dbo.Split(#PermitIDs) -- split function of your choice
UPDATE S
SET
UpdatedOn = GETUTCDATE(),
Status = #Status
FROM
PermitStatus S WITH (UPDLOCK, HOLDLOCK)
INNER JOIN #Permits P ON S.PermitID = P.PermitID
INSERT PermitStatus (
PermitID,
UpdatedOn,
Status
)
SELECT
P.PermitID,
GetUTCDate(),
#Status
FROM #Permits P
WHERE NOT EXISTS (
SELECT 1
FROM PermitStatus S
WHERE P.PermitID = S.PermitID
)
COMMIT TRAN
RETURN ##ERROR;
Note: XACT_ABORT helps guarantee the explicit transaction is closed following a timeout or unexpected error.
To confirm that this handles the locking problem, open several query windows and execute an identical batch like so:
WAITFOR TIME '11:00:00' -- use a time in the near future
EXEC dbo.PermitStatusUpdate #PermitIDs = '123,124,125,126', 1
All of these different sessions will execute the stored procedure in nearly the same instant. Check each session for errors. If none exist, try the same test a few times more (since it's possible to not always have the race condition occur, especially with MERGE).
The writeups at the links I gave above give even more detail than I did here, and also describe what to do for the SQL 2008 MERGE statement as well. Please read those thoroughly to truly understand the issue.
Briefly, with MERGE, no explicit transaction is needed, but you do need to use SET XACT_ABORT ON and use a locking hint:
SET NOCOUNT, XACT_ABORT ON;
MERGE dbo.Table WITH (HOLDLOCK) AS TableAlias
...
This will prevent concurrency race conditions causing errors.
I also recommend that you do error handling after each data modification statement.
If you're using SQL Server 2008, you can use table valued parameters - you pass in a table of records into a stored procedure and then you can do a MERGE.
Passing in a table valued parameter would remove the need to parse CSV strings.
Edit:
ErikE has raised the point about race conditions, please refer to his answer and linked articles.
If you have SQL Server 2008, you can use MERGE. Here's an article describing this.
You should be able to do your insert and your update as two set based queries.
The code below was based on a data load procedure that I wrote a while ago that took data from a staging table and inserted or updated it into the main table.
I've tried to make it match your example, but you may need to tweak this (and create a table valued UDF to parse your CSV into a table of ids).
-- Update where the join on permitstatus matches
Update
PermitStatus
Set
[UpdatedOn]=GETUTCDATE(),
[Status]=staging.Status
From
PermitStatus status
Join
StagingTable staging
On
staging.PermitId = status.PermitId
-- Insert the new records, based on the Where Not Exists
Insert
PermitStatus(Updatedon, Status, PermitId)
Select (GETUTCDATE(), staging.status, staging.permitId
From
StagingTable staging
Where Not Exists
(
Select 1 from PermitStatus status
Where status.PermitId = staging.PermidId
)
Essentially you have an upsert stored procedure (eg. UpsertSinglePermit)
(like the code you have given above) for dealing with one row.
So the steps I see are to create a new stored procedure (UpsertNPermits) which does
a) Parse input string into n record entries (each record contains permit id and status)
b) Foreach entry in above, invoke UpsertSinglePermit
First off, I want to start by saying I am not an SQL programmer (I'm a C++/Delphi guy), so some of my questions might be really obvious. So pardon my ignorance :o)
I've been charged with writing a script that will update certain tables in a database based on the contents of a CSV file. I have it working it would seem, but I am worried about atomicity for one of the steps:
One of the tables contains only one field - an int which must be incremented each time, but from what I can see is not defined as an identity for some reason. I must create a new row in this table, and insert that row's value into another newly-created row in another table.
This is how I did it (as part of a larger script):
DECLARE #uniqueID INT,
#counter INT,
#maxCount INT
SELECT #maxCount = COUNT(*) FROM tempTable
SET #counter = 1
WHILE (#counter <= #maxCount)
BEGIN
SELECT #uniqueID = MAX(id) FROM uniqueIDTable <----Line 1
INSERT INTO uniqueIDTableVALUES (#uniqueID + 1) <----Line 2
SELECT #uniqueID = #uniqueID + 1
UPDATE TOP(1) tempTable
SET userID = #uniqueID
WHERE userID IS NULL
SET #counter = #counter + 1
END
GO
First of all, am I correct using a "WHILE" construct? I couldn't find a way to achieve this with a simple UPDATE statement.
Second of all, how can I be sure that no other operation will be carried out on the database between Lines 1 and 2 that would insert a value into the uniqueIDTable before I do? Is there a way to "synchronize" operations in SQL Server Express?
Also, keep in mind that I have no control over the database design.
Thanks a lot!
You can do the whole 9 yards in one single statement:
WITH cteUsers AS (
SELECT t.*
, ROW_NUMBER() OVER (ORDER BY userID) as rn
, COALESCE(m.id,0) as max_id
FROM tempTable t WITH(UPDLOCK)
JOIN (
SELECT MAX(id) as id
FROM uniqueIDTable WITH (UPDLOCK)
) as m ON 1=1
WHERE userID IS NULL)
UPDATE cteUsers
SET userID = rn + max_id
OUTPUT INSERTED.userID
INTO uniqueIDTable (id);
You get the MAX(id), lock the uniqueIDTable, compute sequential userIDs for users with NULL userID by using ROW_NUMBER(), update the tempTable and insert the new ids into uniqueIDTable. All in one operation.
For performance you need and index on uniqueIDTable(id) and index on tempTable(userID).
SQL is all about set oriented operations, WHILE loops are the code smell of SQL.
You need a transaction to ensure atomicity and you need to move the select and insert into one statement or do the select with an updlock to prevent two people from running the select at the same time, getting the same value and then trying to insert the same value into the table.
Basically
DECLARE #MaxValTable TABLE (MaxID int)
BEGIN TRANSACTION
BEGIN TRY
INSERT INTO uniqueIDTable VALUES (id)
OUTPUT inserted.id INTO #MaxValTable
SELECT MAX(id) + 1 FROM uniqueIDTable
UPDATE TOP(1) tempTable
SET userID = (SELECT MAXid FROM #MaxValTable)
WHERE userID IS NULL
COMMIT TRANSACTION
END TRY
BEGIN CATCH
ROLLBACK TRANSACTION
RAISERROR 'Error occurred updating tempTable' -- more detail here is good
END CATCH
That said, using an identity would make things far simpler. This is a potential concurrency problem. Is there any way you can change the column to be identity?
Edit: Ensuring that only one connection at a time will be able to insert into the uniqueIDtable. Not going to scale well though.
Edit: Table variable's better than exclusive table lock. If need be, this can be used when inserting users as well.
I've got a stored procedure, which selects 1 record back. the stored procedure could be called from several different applications on different PCs. The idea is that the stored procedure brings back the next record that needs to be processed, and if two applications call the stored proc at the same time, the same record should not be brought back. My query is below, I'm trying to write the query as efficiently as possible (sql 2008). Can it get done more efficiently than this?
CREATE PROCEDURE GetNextUnprocessedRecord
AS
BEGIN
SET NOCOUNT ON;
--ID of record we want to select back
DECLARE #iID BIGINT
-- Find the next processable record, and mark it as dispatched
-- Must be done in a transaction to ensure no other query can get
-- this record between the read and update
BEGIN TRAN
SELECT TOP 1
#iID = [ID]
FROM
--Don't read locked records, only lock the specific record
[MyRecords] WITH (READPAST, ROWLOCK)
WHERE
[Dispatched] is null
ORDER BY
[Received]
--Mark record as picked up for processing
UPDATE
[MyRecords]
SET
[Dispatched] = GETDATE()
WHERE
[ID] = #iID
COMMIT TRAN
--Select back the specific record
SELECT
[ID],
[Data]
FROM
[MyRecords] WITH (NOLOCK, READPAST)
WHERE
[ID] = #iID
END
Using the READPAST locking hint is correct and your SQL looks OK.
I'd add use XLOCK though which is also HOLDLOCK/SERIALIZABLE
...
[MyRecords] WITH (READPAST, ROWLOCK, XLOCK)
...
This means you get the ID, and exclusively lock that row while you carry on and update it.
Edit: add an index on Dispatched and Received columns to make it quicker. If [ID] (I assume it's the PK) is not clustered, INCLUDE [ID]. And filter the index too because it's SQL 2008
You could also use this construct which does it all in one go without XLOCK or HOLDLOCK
UPDATE
MyRecords
SET
--record the row ID
#id = [ID],
--flag doing stuff
[Dispatched] = GETDATE()
WHERE
[ID] = (SELECT TOP 1 [ID] FROM MyRecords WITH (ROWLOCK, READPAST) WHERE Dispatched IS NULL ORDER BY Received)
UPDATE, assign, set in one
You can assign each picker process a unique id, and add columns pickerproc and pickstate to your records. Then
UPDATE MyRecords
SET pickerproc = myproc,
pickstate = 'I' -- for 'I'n process
WHERE Id = (SELECT MAX(Id) FROM MyRecords WHERE pickstate = 'A') -- 'A'vailable
That gets you your record in one atomic step, and you can do the rest of your processing at your leisure. Then you can set pickstate to 'C'omplete', 'E'rror, or whatever when it's resolved.
I think Mitch is referring to another good technique where you create a message-queue table and insert the Ids there. There are several SO threads - search for 'message queue table'.
You can keep MyRecords on a "MEMORY" table for faster processing.