I have a requirement to automatically generate a column's value based on another query's result. Because this column value must be unique, I need to take into consideration concurrent requests. This query needs to generate a unique value for a support ticket generator.
The template for the unique value is CustomerName-Month-Year-SupportTicketForThisMonthCount.
So the script should automatically generate:
AcmeCo-10-2019-1
AcmeCo-10-2019-2
AcmeCo-10-2019-3
and so on as support tickets are created. How can ensure that AcmeCo-10-2019-1 is not generated twice if two support tickets are created at the same time for AcmeCo?
insert into SupportTickets (name)
select concat_ws('-', #CustomerName, #Month, #Year, COUNT())
from SupportTickets
where customerName = #CustomerName
and CreatedDate between #MonthStart and #MonthEnd;
One possibility:
Create a counter table:
create table Counter (
Id int identify(1,1),
Name varchar(64)
Count1 int
)
Name is a unique identifier for the sequence, and in your case name would be CustomerName-Month-Year i.e. you would end up with a row in this table for every Customer/Year/Month combination.
Then write a stored procedure similar to the following to allocate a new sequence number:
create procedure [dbo].[Counter_Next]
(
#Name varchar(64)
, #Value int out -- Value to be used
)
as
begin
set nocount, xact_abort on;
declare #Temp int;
begin tran;
-- Ensure we have an exclusive lock before changing variables
select top 1 1 from dbo.Counter with (tablockx);
set #Value = null; -- if a value is passed in it stuffs us up, so null it
-- Attempt an update and assignment in a single statement
update dbo.[Counter] set
#Value = Count1 = Count1 + 1
where [Name] = #Name;
if ##rowcount = 0 begin
set #Value = 10001; -- Some starting value
-- Create a new record if none exists
insert into dbo.[Counter] ([Name], Count1)
select #Name, #Value;
end;
commit tran;
return 0;
end;
You could look into using a TIME type instead of COUNT() to create unique values. That way it is much less likely to have duplicates. Hope that helps
Related
EDIT: Sequential invoice numbering is the law in multiple countries.
EDIT: Poor variable naming on my part suggested I wanted to use my generated Id as a key. This is not the case. Should have stuck with 'invoiceNumber'.
I have the exact same question as posed here https://stackoverflow.com/a/24196374/1980516
However, since the proposed solution threw a syntax error, I've adapted it to use a cursor.
First, there is the stored procedure that generates a new Nr, for a given Business+Year combination:
CREATE PROCEDURE PROC_NextInvoiceNumber #businessId INT, #year INT, #Nr NVARCHAR(MAX) OUTPUT
AS MERGE INTO InvoiceNextNumbers ini
USING (VALUES (#businessId, #year)) Incoming(BusinessId, Year)
ON Incoming.BusinessId = ini.BusinessId AND Incoming.Year = ini.Year
WHEN MATCHED THEN UPDATE SET ini.Nr = ini.Nr + 1
WHEN NOT MATCHED BY TARGET THEN INSERT (BusinessId, Year, Nr)
VALUES(#businessId, #year, 1)
OUTPUT INSERTED.Nr;
Then, using that stored procedure, I've created an INSTEAD OF INSERT trigger:
CREATE TRIGGER TRIG_GenerateInvoiceNumber ON Invoices INSTEAD OF INSERT
AS
BEGIN
DECLARE #BusinessId INT
DECLARE #InvoiceId INT
DECLARE #BillingDate DATETIME2(7)
-- Cursors are expensive, but I don't see any other way to call the stored procedure per row
-- Mitigating factor: Mostly, we're only inserting one Invoice at a time
DECLARE InsertCursor CURSOR FAST_FORWARD FOR
SELECT BusinessId, Id, BillingDate FROM INSERTED
OPEN InsertCursor
FETCH NEXT FROM InsertCursor
INTO #BusinessId, #InvoiceId, #BillingDate
WHILE ##FETCH_STATUS = 0
BEGIN
DECLARE #year INT
SET #year = year(#BillingDate)
DECLARE #Number NVARCHAR(MAX)
EXEC PROC_NextInvoiceNumber #BusinessId, #year, #Number OUTPUT
-- SET #Number = 'this works'
INSERT INTO Invoices (BusinessId, Id, BillingDate, Number)
VALUES (#BusinessId, #InvoiceId, #BillingDate, #Number)
FETCH NEXT FROM InsertCursor
INTO #BusinessId, #InvoiceId, #BillingDate
END
CLOSE InsertCursor
DEALLOCATE InsertCursor
END
If I uncomment SET #Number = 'this works', then in my database that exact string ('this works') is successfully set in Invoice.Number.
Somehow, my OUTPUT parameter is not set and I can't figure out why not.. Can someone shed a light on this?
EDIT update in response to comments (thank you):
I have a composite key (BusinessId, Id) for Invoice. The desired end result is a unique Invoice Identifier Number of the form '20180001' that is a continuous sequence of numbers within the businessId. So business 1 has invoice Numbers 20180001, 20180002, 20180003 and business 2 also has invoice numbers 20180001, 20180002, 20180003. (But different composite primary keys)
I don't want that cursor either, but I saw no other way within the framework as suggested by the question I refer to up above.
Manual call of PROC_NextInvoiceNumber with existing business id and year returns NULL.
If I try to set Id in PROC_NextInvoiceNumber, I get A MERGE statement must be terminated by a semi-colon (;). if I set it inside the MERGE or The multi-part identifier "INSERTED.Nr" could not be bound. if I set outside the MERGE.
Your OUTPUT parameter is never set. You are using the OUTPUT clause of the MERGE statement to create a result set. This is unrelated to assigning a value to a parameter.
MERGE INTO..
USING ... ON ...
WHEN MATCHED THEN UPDATE ...
WHEN NOT MATCHED BY TARGET THEN INSERT ...
OUTPUT INSERTED.Nr; /* <-- HERE this is the OUTPUT *clause* */
Change the code to actually assign something to #Nr:
SET #Nr = ...
The typical way is to use the OUTPUT clause to store the desired value into a table variable and then assign the value to the desired output *variable:
DECLARE #t TABLE (Nr NVARCHAR(MAX));
MERGE INTO..
USING ... ON ...
WHEN MATCHED THEN UPDATE ...
WHEN NOT MATCHED BY TARGET THEN INSERT ...
OUTPUT INSERTED.Nr INTO #t;
SELECT #Nr = Nr FROM #t;
I have a table with 5 billions of rows in SQL Server 2014 (Developer Edition, x64, Windows 10 Pro x64):
CREATE TABLE TestTable
(
ID BIGINT IDENTITY(1,1),
PARENT_ID BIGINT NOT NULL,
CONSTRAINT PK_TestTable PRIMARY KEY CLUSTERED (ID)
);
CREATE NONCLUSTERED INDEX IX_TestTable_ParentId
ON TestTable (PARENT_ID);
I'm trying to apply the following patch:
-- Create non-nullable column with default (should be online operation in Enterprise/Developer edition)
ALTER TABLE TestTable
ADD ORDINAL TINYINT NOT NULL CONSTRAINT DF_TestTable_Ordinal DEFAULT 0;
GO
-- Populate column value for existing data
BEGIN
SET NOCOUNT ON;
DECLARE #BATCH_SIZE BIGINT = 1000000;
DECLARE #COUNTER BIGINT = 0;
DECLARE #ROW_ID BIGINT;
DECLARE #ORDINAL BIGINT;
DECLARE ROWS_C CURSOR
LOCAL FORWARD_ONLY FAST_FORWARD READ_ONLY
FOR
SELECT
ID AS ID,
ROW_NUMBER() OVER (PARTITION BY PARENT_ID ORDER BY ID ASC) AS ORDINAL
FROM
TestTable;
OPEN ROWS_C;
FETCH NEXT FROM ROWS_C
INTO #ROW_ID, #ORDINAL;
BEGIN TRANSACTION;
WHILE ##FETCH_STATUS = 0
BEGIN
UPDATE TestTable
SET
ORDINAL = CAST(#ORDINAL AS TINYINT)
WHERE
ID = #ROW_ID;
FETCH NEXT FROM ROWS_C
INTO #ROW_ID, #ORDINAL;
SET #COUNTER = #COUNTER + 1;
IF #COUNTER = #BATCH_SIZE
BEGIN
COMMIT TRANSACTION;
SET #COUNTER = 0;
BEGIN TRANSACTION;
END;
END;
COMMIT TRANSACTION;
CLOSE ROWS_C;
DEALLOCATE ROWS_C;
SET NOCOUNT OFF;
END;
GO
-- Drop default constraint from the column
ALTER TABLE TestTable
DROP CONSTRAINT DF_TestTable_Ordinal;
GO
-- Drop IX_TestTable_ParentId index
DROP INDEX IX_TestTable_ParentId
ON TestTable;
GO
-- Create IX_TestTable_ParentId_Ordinal index
CREATE UNIQUE INDEX IX_TestTable_ParentId_Ordinal
ON TestTable (PARENT_ID, ORDINAL);
GO
The aim of patch is to add a column, called ORDINAL, which is an ordinal number of the record within the same parent (defined by PARENT_ID). The patch is run using SQLCMD.
The patch is done is this way for a set of reasons:
Table is too large to run a single UPDATE statement on it (takes enormous amount of time and space in transaction log/tempdb).
Batch updates using a single UPDATE statement with TOP n rows are not simple to implement (if we update table in, say, 1m rows batches, 1000001st row may belong to the same PARENT_ID as 1000000th which will lead to wrong ordinal number assigned to 1000001st record). In other words, SELECT statement run in cursor should be run once (without paging) or more complicated operations (joins/conditions) should be applied.
Adding NULL column and changing it to NOT NULL later is not a good solution since I use SNAPSHOT isolation (full table update will be performed on altering column to be NOT NULL).
The patch works perfect on a small database with a few millions of rows, but, when applied to the one with billions of rows, I get:
Msg 3606, Level 16, State 2, Server XXX, Line 22
Arithmetic overflow occurred.
My first guess was ORDINAL value is too big to fit into TINYINT column, but this is not the case. I created a test database with similar structure and populated with data (more than 255 rows per parent). The error message I get is still arithmetic exception, but with different message code and different wording (explicitly saying it can't fit data into TINYINT).
Currently I have a couple of suspicions, but I haven't managed to find anything that could help me:
CURSOR is not able to handle more than MAX(INT32) rows.
SQLCMD imposed limitations.
Do you have any ideas on what could the problem be?
How about using a While loop but making sure that you keep the same parent_ids together:
DECLARE #SegmentSize BIGINT = 1000000
DECLARE #CurrentSegment BigInt = 0
WHILE 1 = 1
BEGIN
;With UpdateData As
(
SELECT ID AS ID,
ROW_NUMBER() OVER (PARTITION BY PARENT_ID ORDER BY ID ASC) AS ORDINAL
FROM TestData
WHERE ID > #CurrentSegment AND ID <= (#CurrentSegment + #SegmentSize)
)
UPDATE TestData
SET Ordinal = UpdateDate.Ordinal
FROM TestData
INNER JOIN UpdateData ON TestData.Id = UpdateData.Id
IF ##ROWCOUNT = 0
BEGIN
BREAK
END
SET #CurrentSegment = #CuurentSegment + #SegmentSize
END
EDIT - Amended to segment on Parent_Id as per request. This should be
reasonably quick as Parent_id is indexed (added Option(Recompile)
to ensure that actual value is used for the lookup.
Because you are not updating
the whole table this will limit the Transaction Log growth!
DECLARE #SegmentSize BIGINT = 1000000
DECLARE #CurrentSegment BigInt = 0
WHILE 1 = 1
BEGIN
;With UpdateData As
(
SELECT ID AS ID,
ROW_NUMBER() OVER (PARTITION BY PARENT_ID ORDER BY ID ASC) AS ORDINAL
FROM TestData
WHERE Parent_ID > #CurrentSegment AND
Parent_ID <= (#CurrentSegment + #SegmentSize)
)
UPDATE TestData
SET Ordinal = UpdateDate.Ordinal
FROM TestData
INNER JOIN UpdateData ON TestData.Id = UpdateData.Id
OPTION (RECOMPILE)
IF ##ROWCOUNT = 0
BEGIN
BREAK
END
SET #CurrentSegment = #CuurentSegment + #SegmentSize
END
I'm writing a DML trigger when change (update or Insert) happens in one table (Master table), I want to write the whole row into another table (Delta table).
The Master and Delta tables contains the same column with same datatype, except that Delta table contains an additional column called 'change_type', which should say either 'INSERT' OR 'MODIFY', depending on which trigger is updating the delta table.
The difficulty I'm having is I want to use the inserted table to update the Delta table row but its giving me errors.
CREATE TRIGGER [dbo].[TR_Update]
ON [dbo].[People_Master]
AFTER Update
AS
BEGIN
-- SET NOCOUNT ON added to prevent extra result sets from
-- interfering with SELECT statements.
SET NOCOUNT ON;
Declare #RowCount int
Declare #ID int
Declare #Email nvarchar(50)
Declare #ct nvarchar(10)
select #ID = ID from inserted
Select #RowCount=COUNT(*) from People_Delta where People_Delta.ID = #ID and People_Delta.change_type = 'Modify';
if(#RowCount = 0)
Begin
Insert into People_Delta (ID,Email,uac,Department,FirstName,change_type)
values (iserted.ID,inserted.Email,inserted.uac,inserted.Department,inserted.Firstname'Modify');
END
END
GO
My table has 5 columns.
ID (primary key)
Email
Firstname
uac
Department
You are missing a , in your INSERT statement.
And because the number of columns you have specified does not match with the number of values you are inserting, you get an error.
inserted.Firstname , 'Modify'
Insert into People_Delta (ID,Email,uac,Department,FirstName,change_type)
values (iserted.ID,inserted.Email,inserted.uac,inserted.Department,inserted.Firstname,'Modify');
I have a table called Employee. The EmpId column serves as the primary key. In my scenario, I cannot make it AutoNumber.
What would be the best way of generating the the next EmpId for the new row that I want to insert in the table?
I am using SQL Server 2008 with C#.
Here is the code that i am currently getting, but to enter Id's in key value pair tables or link tables (m*n relations)
Create PROCEDURE [dbo].[mSP_GetNEXTID]
#NEXTID int out,
#TABLENAME varchar(100),
#UPDATE CHAR(1) = NULL
AS
BEGIN
DECLARE #QUERY VARCHAR(500)
BEGIN
IF EXISTS (SELECT LASTID FROM LASTIDS WHERE TABLENAME = #TABLENAME and active=1)
BEGIN
SELECT #NEXTID = LASTID FROM LASTIDS WHERE TABLENAME = #TABLENAME and active=1
IF(#UPDATE IS NULL OR #UPDATE = '')
BEGIN
UPDATE LASTIDS
SET LASTID = LASTID + 1
WHERE TABLENAME = #TABLENAME
and active=1
END
END
ELSE
BEGIN
SET #NEXTID = 1
INSERT INTO LASTIDS(LASTID,TABLENAME, ACTIVE)
VALUES(#NEXTID+1,#TABLENAME, 1)
END
END
END
Using MAX(id) + 1 is a bad idea both performance and concurrency wise.
Instead you should resort to sequences which were design specifically for this kind of problem.
CREATE SEQUENCE EmpIdSeq AS bigint
START WITH 1
INCREMENT BY 1;
And to generate the next id use:
SELECT NEXT VALUE FOR EmpIdSeq;
You can use the generated value in a insert statement:
INSERT Emp (EmpId, X, Y)
VALUES (NEXT VALUE FOR EmpIdSeq, 'x', 'y');
And even use it as default for your column:
CREATE TABLE Emp
(
EmpId bigint PRIMARY KEY CLUSTERED
DEFAULT (NEXT VALUE FOR EmpIdSeq),
X nvarchar(255) NULL,
Y nvarchar(255) NULL
);
Update: The above solution is only applicable to SQL Server 2012+. For older versions you can simulate the sequence behavior using dummy tables with identity fields:
CREATE TABLE EmpIdSeq (
SeqID bigint IDENTITY PRIMARY KEY CLUSTERED
);
And procedures that emulates NEXT VALUE:
CREATE PROCEDURE GetNewSeqVal_Emp
#NewSeqVal bigint OUTPUT
AS
BEGIN
SET NOCOUNT ON
INSERT EmpIdSeq DEFAULT VALUES
SET #NewSeqVal = scope_identity()
DELETE FROM EmpIdSeq WITH (READPAST)
END;
Usage exemple:
DECLARE #NewSeqVal bigint
EXEC GetNewSeqVal_Emp #NewSeqVal OUTPUT
The performance overhead of deleting the last inserted element will be minimal; still, as pointed out by the original author, you can optionally remove the delete statement and schedule a maintenance job to delete the table contents off-hour (trading space for performance).
Adapted from SQL Server Customer Advisory Team Blog.
Working SQL Fiddle
The above
select max(empid) + 1 from employee
is the way to get the next number, but if there are multiple user inserting into the database, then context switching might cause two users to get the same value for empid and then add 1 to each and then end up with repeat ids. If you do have multiple users, you may have to lock the table while inserting. This is not the best practice and that is why the auto increment exists for database tables.
I hope this works for you. Considering that your ID field is an integer
INSERT INTO Table WITH (TABLOCK)
(SELECT CASE WHEN MAX(ID) IS NULL
THEN 1 ELSE MAX(ID)+1 END FROM Table), VALUE_1, VALUE_2....
Try following query
INSERT INTO Table VALUES
((SELECT isnull(MAX(ID),0)+1 FROM Table), VALUE_1, VALUE_2....)
you have to check isnull in on max values otherwise it will return null in final result when table contain no rows .
trying to get this stored procedure to work.
ALTER PROCEDURE [team1].[add_testimonial]
-- Add the parameters for the stored procedure here
#currentTestimonialDate char(10),#currentTestimonialContent varchar(512),#currentTestimonialOriginator varchar(20)
AS
BEGIN
DECLARE
#keyValue int
SET NOCOUNT ON;
--Get the Highest Key Value
SELECT #keyValue=max(TestimonialKey)
FROM Testimonial
--Update the Key by 1
SET #keyValue=#keyValue+1
--Store into table
INSERT INTO Testimonial VALUES (#keyValue, #currentTestimonialDate, #currentTestimonialContent, #currentTestimonialOriginator)
END
yet it just returns
Running [team1].[add_testimonial] ( #currentTestimonialDate = 11/11/10, #currentTestimonialContent = this is a test, #currentTestimonialOriginator = theman ).
No rows affected.
(0 row(s) returned)
#RETURN_VALUE = 0
Finished running [team1].[add_testimonial].
and nothing is added to the database, what might be the problem?
There may have problems in two place:
a. There is no data in the table so, max(TestimonialKey) returns null, below is the appropriate way to handle it.
--Get the Highest Key Value
SELECT #keyValue= ISNULL(MAX(TestimonialKey), 0)
FROM Testimonial
--Update the Key by 1
SET #keyValue=#keyValue+1
b. Check your data type of the column currentTestimonialDate whether it is char or DateTime type, if this field is datetime type in the table then convert #currentTestimonialDate to DateTime before inserting to the table.
Also, check number of columns that are not null allowed and you're passing data to them.
If you're not passing data for all columns then try by specifying columns name as below:
--Store into table
INSERT INTO Testimonial(keyValue, currentTestimonialDate,
currentTestimonialContent, currentTestimonialOriginator)
VALUES (#keyValue, #currentTestimonialDate,
#currentTestimonialContent, #currentTestimonialOriginator)
EDIT:
After getting the comment from marc_s:
Make keyValue as INT IDENTITY, If multiple user call it concurrently that wont be problem, DBMS will handle it, so the ultimate query in procedure might be as below:
ALTER PROCEDURE [team1].[add_testimonial]
-- Add the parameters for the stored procedure here
#currentTestimonialDate char(10),
#currentTestimonialContent varchar(512),#currentTestimonialOriginator varchar(20)
AS
BEGIN
SET NOCOUNT ON;
--Store into table
INSERT INTO Testimonial VALUES (#currentTestimonialDate,
#currentTestimonialContent, #currentTestimonialOriginator)
END
Two issues that I can spot:
SELECT #keyValue=max(TestimonialKey)
should be
SELECT #keyValue=ISNULL(max(TestimonialKey), 0)
To account for the case when there are no records in the database
Second, I believe that with NOCOUNT ON, you will not return the count of inserted rows to the caller. So, before your INSERT statement, add
SET NOCOUNT OFF