Friday, August 17, 2012

SQL Query : Update row columns with another row

How to update row columns with another row if column values are null in SQL

Update row column with another row column in same table SQL only if it is null.

Background

To test queries and concept I am using SQL SERVER 2008 R2.

Queries

If you need to update one row columns from another row columns in the same table following approach can be useful.

Create sample table.
CREATE TABLE [dbo].[Employee](
      [Title] [varchar](50) NULL,
      [FirstName] [varchar](50) NULL,
      [Surname] [varchar](50) NULL,
      [DOB] [varchar](50) NULL,
      [Datestarted] [varchar](50) NULL,
      [EmployeeId] [int] IDENTITY(1,1) NOT NULL,
      [Interest] [varchar](50) NULL,
      [EmpAddress] [varchar](500) NULL
) ON [PRIMARY]
GO
Add some sample rows as follows.

I am just updating 2 columns. Of course you can update more in similar fashion. Here updating 2 columns of 1st row with 2nd row column values.
UPDATE n
   SET n.EmpAddress = p.EmpAddress
      ,n.[Interest] = p.Interest
      FROM [Employee] n inner join [Employee] p
ON n.employeeid = 1 AND p.employeeid = 2
GO

Select * from Employee


Please see next query if you wish to update only when value in the column is null.

Update column when value is null

I am going to use COALESCE. This is function similar to Case.
It returns the first nonnull expression among its arguments. Example
select coalesce(Interest,'FUN') from [Employee] where EmployeeId = 5

Updating two columns of row 4th when they are null with column values of row 3rd.
Precise query can look like
UPDATE n
   SET n.EmpAddress = coalesce(n.EmpAddress, p.EmpAddress)
      ,n.[Interest] = coalesce(n.Interest, p.Interest)
      FROM [Employee] n inner join [Employee] p
ON n.employeeid = 4 AND p.employeeid = 3
GO

Select * from Employee

Summary

In this article I have suggested queries to update row with anther row and how you can use coalesce to restrict modification to null fields only. I hope you can use these queries for other purposes too.


Tag: SQL, SQL SERVER, SQL Server 2008, COALESCE, Update column when value is null, How to update row with another row in SQL, update row with another row column in same table SQL, Update table column with data from other columns in same row, SQL UPDATE with sub-query that references the same table, SQL UPDATE from another row in the same table, Update column value based on other columns in same table, Update data in one table with data from another table, update when value is null sql, How can I update a value on a column only if that value is null, update a column value in case null, update a column value when null, SQL UPDATE, but only if the old value is null, Update columns if it is a null, Update only if null, IS NULL and IS NOT NULL, Update to 0 if NULL

Tuesday, August 7, 2012

SQL Server : Data sharing across stored procedures

Table data sharing across stored procedures


This article demonstrates how table data can be shared in two stored procedures. There are several requirements where we need to pass temp table data to called stored procedure for further processing and can return manipulated data to parent stored procedure.
Background
To test queries and concept I am using SQL SERVER 2008 R2.
Introduction
I want to evaluate some of the good ways to accomplish this. I shall test approach for concurrent calling.
Table Valued Function – This can be good approach. But limitation is you can’t call stored procedure inside function if you need to.
Using Temp Table - This approach looks promising. It works for in and out.
Passing table variable – By any chance if you are using insert into in both parent and child stored procedures. Then it fails.
Other approaches like Using Cursor Variables, CLR, Open query or XML are complex, non efficient or having other pitfalls.

Evaluation of data sharing using temp table

Lets evaluate most promising approach. Test - whether temp tables are call dependent and not, causing problem in other calls.
Create table named as mytable.
CREATE TABLE [dbo].[MyTable](
            [col1] [int] NOT NULL,
            [col2] [char](5) NULL
) ON [PRIMARY]
Lets insert some sample rows for test.
INSERT INTO [MyTable]
            ([col1],
             [col2])
VALUES      (1,
             A)
Go
INSERT INTO [MyTable]
            ([col1],
             [col2])
VALUES      (2,
             B)
GO
INSERT INTO [MyTable]
            ([col1],
             [col2])
VALUES      (3,
             C)
Go
INSERT INTO [MyTable]
            ([col1],
             [col2])
VALUES      (4,
             D)
Go


Create child stored procedure named as called_procedure. Here first I am checking whether temp table is existing or not. If exists then insert some data as per passed parameter @par1.
-- If Exist then drop and create
IF EXISTS (SELECT *
           FROM   sys.objects
           WHERE  object_id = Object_id(N'[dbo].[called_procedure]')
                  AND type IN ( N'P', N'PC' ))
  DROP PROCEDURE [dbo].[called_procedure]

GO

CREATE PROCEDURE Called_procedure @par1 INT,
                                  @par2 BIT
AS
  BEGIN
      IF Object_id('tempdb..#mytemp') IS NOT NULL
        BEGIN
            INSERT INTO #mytemp
            SELECT *
            FROM   Mytable
            WHERE  col1 = @par1
        END
  END

Go
Create caller Stored procedure without creating temp table in scope.
IF EXISTS (SELECT *
           FROM   sys.objects
           WHERE  object_id = Object_id(N'[dbo].[caller_procedure1]')
                  AND type IN ( N'P', N'PC' ))
  DROP PROCEDURE [dbo].[caller_procedure1]

GO

CREATE PROCEDURE Caller_procedure1
AS
  BEGIN
      --Testing for if temp table does not exists in scope
      EXEC Called_procedure
        1,
        0

      IF Object_id('tempdb..#mytemp') IS NOT NULL
        BEGIN
            SELECT *
            FROM   #mytemp
        END
  END

GO
Another caller stored procedure with temp table. This procedure can be called for different parameters.
IF EXISTS (SELECT *
           FROM   sys.objects
           WHERE  object_id = Object_id(N'[dbo].[caller_procedure2]')
                  AND type IN ( N'P', N'PC' ))
  DROP PROCEDURE [dbo].[caller_procedure2]

GO

CREATE PROCEDURE Caller_procedure2 @par1 INT
AS
  BEGIN
      CREATE TABLE #mytemp
        (
           col1 INT NOT NULL,
           col2 CHAR(5) NULL
        )

      EXEC Called_procedure
        @par1,
        0

      SELECT *
      FROM   #mytemp
  END

go

Execute all following queries same time. You can also execute these procedures same time from different systems.
CREATE TABLE #mytemp (col1 int     NOT NULL,
                         col2 char(5) NULL
                        )
Exec caller_procedure2 2
Exec caller_procedure2 4
Exec caller_procedure2 2
Exec caller_procedure2 4
drop table #mytemp
Exec caller_procedure1


Caller_procedure2 takes it’s own temp table and drops when scope ends. It does not take temp table which we have created outside the proc.

Summary

This little interesting exercise can help you to understand how you can pass temp table to child stored procedures. Using this concept you can write multipurpose stored procedures to increase reusability of code.
If this article helps you in designing/coding SQL logic don’t forget to hit voting option. Please comment your suggestions and improvements.
Happy Coding!!

Search Tags:
SQL, DML, DDL, SQL SERVER 2008, SQL SERVER 2005, SQL SERVER 2000, SP, Stored Procedure, Function, Temp Table, Table Variable, Passing temp table to Stored Procedure, Table data sharing across stored procedures, Data sharing in SQL, Writing reusable Stored Procedures, Passing Rows to Stored Procedures