zlika's » MS SQL Server

How to store UTF-8 data in SQL Server

zlika — Sun, 06 Mar 2011 22:09:02 +0000

Very frequently we need to use the data type UTF-8 to store a data in different languages. As described in the MSDN page the UTF-8 data type is stored in the ‘Samples’ folder in SQL Server (Server\100\Samples\Engine\Programmability\CLR\UTF8String). You just need to activate it.

Create data type UTF-8

1. Generate a strong name key file
1.1. Open Command Prompt (Start –> Run –> ‘cmd’)
1.2. Change directory (C:\Program Files\Microsoft SDKs\Windows\v7.0A\bin)
1.3. Generate the key (C:\Program Files\Microsoft SDKs\Windows\v7.0A\bin>sn -k “C:\Program Files\Microsoft SQL Server\100\Samples\SampleKey.snk”)

2. Build the sample
2.1. Open Microsoft Visual Studio
2.2. Open ‘C:\Program Files\Microsoft SQL Server\100\Samples\Engine\Programmability\CLR\UTF8String\CS\UTF8String.sln’
2.3. Build –> Build UTF8String

3. Install UTF-8 data type
3.1. Open SQL Server Management Studio
3.2. Open ‘C:\Program Files\Microsoft SQL Server\100\Samples\Engine\Programmability\CLR\UTF8String\Scripts\InstallCS.sql’
3.3. Execute it
3.4. Test UTF-8 data type (C:\Program Files\Microsoft SQL Server\100\Samples\Engine\Programmability\CLR\UTF8String\Scripts\Test.sql)
3.4.1. Fix the error message ‘Msg 6263, Level 16, State 1, Line 2
Execution of user code in the .NET Framework is disabled. Enable “clr enabled” configuration option.‘

sp_configure 'clr enabled', 1;
RECONFIGURE;

3.4.2. Test UTF-8

DECLARE @u Utf8String;
SET @u = CONVERT(Utf8String, 'hello world');
SELECT
	@u.ToString()
	, CONVERT(varbinary(8000), @u)
	, SUBSTRING(@u.ToString(), 1, 5)
	, @u.Utf8Bytes;

The result

Create PHP form to insert a UTF-8 data

1. Create table

USE zlika;

CREATE TABLE CharsetTest
(
	[Language] VARCHAR(50)
	, [Phrase] UTF8String
);

2. Create PHP form (insert.php)



	
	Charset Test






	Language: 

	Phrase:

3. Insert test data

Bulgarian –> Честита Нова Година!
Arabic –> سنة جديدة سعيدة!
Armenian — > Շնորհավոր Նոր Տարի
Azarbaijani –> Yeni iliniz mübarək
Catalan –> Feliç Any Nou!
Chinese (Traditional) –> 新年快樂
Polish –> Szczęśliwego Nowego Roku!
Thai –> สวัสดีปีใหม่
hindi –> नया साल मुबारक हो

4. Check the content of the table (check.php)



	
	Charset Test



';

$i = 0;
while ($i < mssql_num_fields($result))
{
	echo ''. mssql_field_name($result, $i) . '';
	$i++;
}
echo '';

//Retrieving rows
while ($row = mssql_fetch_array($result, MSSQL_ASSOC))
{
	echo '';
	foreach ($row as $data)
	{
		echo '' . $data . '';
	}
	echo'';
}

echo '';

// Close database connection
mssql_close($connect);

?>

The result:

The result

Félicitations

Feed enhanced by Better Feed from Ozh

MaraK

zlika — Thu, 17 Feb 2011 16:41:01 +0000

I’m starting a project for management of a small store, using C# and SQL Server Express.

Félicitations

Feed enhanced by Better Feed from Ozh

SELECT * or SELECT [Column Name] in IF statement in MS SQL Server

zlika — Sun, 31 Oct 2010 19:02:32 +0000

When we decide how to manipulate the data, we use IF statement verification. For example if we export information about merchant sales, we have two common groups of merchants – these we scan (with serial numbers) and these we don’t scan. For both of them we use two BEGIN… END blocks. To verify the type we use IF statement which searches for a row in ‘RegisteredScans’ table.

We can write the IF statement like:

...
IF EXISTS
(
	SELECT *
	FROM RegisteredScans
	WHERE...
)
BEGIN
	...
END
...

or:

...
IF EXISTS
(
	SELECT [Column Name]
	FROM RegisteredScans
	WHERE [Column Name]...
)
BEGIN
	...
END
...

The difference is that we select ALL or only the column that we use to verify the truth of the IF statement.

I created an example to demonstrate the difference between SELECT * and SELECT [Column Name]. First I need to know which table has most rows in AdventureWorks2008 database. On this address I found the following script:

USE AdventureWorks2008
GO

ALTER PROCEDURE dbo.listTableRowCounts
AS
BEGIN
	SET NOCOUNT ON

	DECLARE @SQL VARCHAR(255)
	SET @SQL = 'DBCC UPDATEUSAGE (' + DB_NAME() + ')'
	EXEC(@SQL)

	CREATE TABLE #foo
	(
		tablename VARCHAR(255)
        , rc INT
	)

	INSERT #foo
	EXEC sp_msForEachTable
		'SELECT PARSENAME(''?'', 1),
		COUNT(*) FROM ?'
END

	SELECT tablename, rc
	FROM #foo
	ORDER BY rc DESC

	DROP TABLE #foo
GO

Then I executed the stored procedure

USE AdventureWorks2008
GO

EXEC dbo.listTableRowCounts
GO

And I found that the table ‘SalesOrderDetail’ contains 121317 rows (most in the database). Next I wrote a script which uses Start time and End time to calculate the duration of execution of a code. In our example the code is:

SELECT *
FROM SalesOrderDetail
GO

and

SELECT SalesOrderID
FROM SalesOrderDetail
GO

When I select all the columns, the time for the operation is 6.594 seconds

USE AdventureWorks2008
GO

-- Declare variables
DECLARE @Start TIME
DECLARE @End TIME

-- Set start time
SET @Start = GETDATE()

-- Operation
SELECT *
FROM Sales.SalesOrderDetail

-- Set end time
SET @End = GETDATE()

-- Calculate the difference in times
SELECT DATEDIFF(MS, @Start, @End) AS Milliseconds
GO

SELECT * FROM...

When I select the ‘SalesOrderID’ column, the time for the operation is 2.470 seconds

SELECT SalesOrderID FROM...

When I select ‘rowguid’ column, the time for the operation is 3.234 seconds

SELECT rowguid FROM...

The difference between ‘SalesOrderID’ and ‘rowguid’ columns is that ‘SalesOrderID’ is Primary Key, i.e. it’s indexed.

Félicitations

Feed enhanced by Better Feed from Ozh

Select server logins and their roles in SQL Server

zlika — Sun, 24 Oct 2010 21:25:37 +0000

To view all the server logins and their roles I use the system view sys.syslogins and CASE to convert 1 to “Yes” and 0 to “No”.

-- Select all database logins and their server roles
USE master
GO

SELECT
	[name] AS [Login]
	, CASE sysadmin WHEN 0 THEN 'No' WHEN 1 THEN 'Yes' ELSE 'N/A' END AS [sysadmin]
	, CASE securityadmin WHEN 0 THEN 'No' WHEN 1 THEN 'Yes' ELSE 'N/A' END AS [securityadmin]
	, CASE serveradmin WHEN 0 THEN 'No' WHEN 1 THEN 'Yes' ELSE 'N/A' END AS [serveradmin]
	, CASE setupadmin WHEN 0 THEN 'No' WHEN 1 THEN 'Yes' ELSE 'N/A' END AS [setupadmin]
	, CASE processadmin WHEN 0 THEN 'No' WHEN 1 THEN 'Yes' ELSE 'N/A' END AS [processadmin]
	, CASE diskadmin WHEN 0 THEN 'No' WHEN 1 THEN 'Yes' ELSE 'N/A' END AS [diskadmin]
	, CASE dbcreator WHEN 0 THEN 'No' WHEN 1 THEN 'Yes' ELSE 'N/A' END AS [dbcreator]

FROM sys.syslogins
GO

Félicitations

Feed enhanced by Better Feed from Ozh

Filegroups and Autogrowth in MS SQL Server

zlika — Sun, 19 Sep 2010 03:27:10 +0000

Yesterday I was asked about the way a table is stored in several files in a filegroup.

The current example explains it. First, I create a database, containing two secondary data files in a filegroup.

CREATE DATABASE zlika ON PRIMARY
(
	NAME = N'zlika'
	, FILENAME = N'C:\Databases\zlika.mdf'
)
,

FILEGROUP zlikaGroup
(
	NAME = N'fg1_1'
	, FILENAME = N'C:\Databases\fg1_1.ndf'
)
,
(
	NAME = N'fg1_2'
	, FILENAME = N'C:\Databases\fg1_2.ndf'
)

LOG ON
(
	NAME = N'zlika_log'
	, FILENAME = N'C:\Databases\zlika_log.ldf'
)
GO

The sizes of the files are as follow:

File sizes after creation of the database

Next I create a table that stores its rows in the filegroup

USE zlika
GO

CREATE TABLE TestTable
(
	ID INT IDENTITY NOT NULL
	, FirstName VARCHAR(50) NOT NULL
	, LastName VARCHAR(50) NOT NULL
)
ON zlikaGroup
GO

I populate the table with any information

USE zlika
GO

INSERT INTO TestTable VALUES ('George', 'Henderson')
GO 50000

The sizes of the files are as follow:

File sizes after the population of the table

I execute the same script and check the sizes of the files once again:

File sizes after adding more data

The answer is: MS SQL Server stores the data in several files in a filegroup according to their Autogrowth settings. The database I have created autogrows the secondary data files by 1MB:

The properties of the database - Files

Félicitations

Feed enhanced by Better Feed from Ozh

DENSE_RANK() function in MS SQL Server

zlika — Thu, 05 Aug 2010 04:46:24 +0000

As we can read in Books Online in SQL Server

Ranking functions return a ranking value for each row in a partition. Depending on the function that is used, some rows might receive the same value as other rows. Ranking functions are nondeterministic.

I have created simple example for usage of the Ranking function DENSE_RANK().

We have a database with the addresses of all the residents of a city. Many of them have changed their addresses several times. They have changed them by using web application, because the authorities have decided that’s the most convenient manner for everybody. Down the years, one user may have more then one record (change of address). We need to select only one actual address or only one precedent address per user. We need to partition the table and select only the data we need.

First I create the tables, keys and I populate them.

USE zlika
GO

-- Create Tables and keys
CREATE TABLE [Users]
(
	[ID] INT IDENTITY (1, 1) PRIMARY KEY
	, [FirstName] VARCHAR(50)
	, [LastName] VARCHAR(50)
)
GO

CREATE TABLE [Address]
(
	[ID] INT IDENTITY (1, 1) PRIMARY KEY
	, [UserID] INT
	, [StreetNo] INT
	, [StreetName] VARCHAR(100)
	, [City] VARCHAR(100)
	, [Phone] VARCHAR(25)
	, [DateOfRegistration] DATETIME NULL
)
GO

ALTER TABLE [Address]
WITH CHECK ADD CONSTRAINT
FK_Users_ID FOREIGN KEY
(
	UserID
)
REFERENCES Users
(
	ID
)
GO

ALTER TABLE [Address]
ADD CONSTRAINT DF_Address_DateOfRegistration
	DEFAULT (GETDATE())
	FOR DateOfRegistration
GO

-- Populate Tables
INSERT INTO [Users] (FirstName, LastName) VALUES
('First', 'Smith')
, ('Second', 'George')
, ('Third', 'Thompson')
GO

SELECT * FROM Users
GO

INSERT INTO [Address] (UserID, StreetNo, StreetName, City, Phone) VALUES (1, 111, 'StreetName1', 'City1', 11111)
WAITFOR DELAY '00:00:01'
INSERT INTO [Address] (UserID, StreetNo, StreetName, City, Phone) VALUES (1, 222, 'StreetName2', 'City2', 22222)
WAITFOR DELAY '00:00:01'
INSERT INTO [Address] (UserID, StreetNo, StreetName, City, Phone) VALUES (1, 333, 'StreetName3', 'City3', 33333)
WAITFOR DELAY '00:00:01'
INSERT INTO [Address] (UserID, StreetNo, StreetName, City, Phone) VALUES (1, 444, 'StreetName4', 'City4', 44444)
WAITFOR DELAY '00:00:01'
INSERT INTO [Address] (UserID, StreetNo, StreetName, City, Phone) VALUES (1, 555, 'StreetName5', 'City5', 55555)
WAITFOR DELAY '00:00:01'
INSERT INTO [Address] (UserID, StreetNo, StreetName, City, Phone) VALUES (2, 1111, 'StreetName11', 'City11', 1111111111)
WAITFOR DELAY '00:00:01'
INSERT INTO [Address] (UserID, StreetNo, StreetName, City, Phone) VALUES (2, 2222, 'StreetName22', 'City22', 2222222222)
WAITFOR DELAY '00:00:01'
INSERT INTO [Address] (UserID, StreetNo, StreetName, City, Phone) VALUES (3, 3, 'StreetName333', 'City333', 333)
GO

SELECT * FROM [Address]
GO

Users table contains three rows as follow:

Users Table

Address table contains eight rows for these three users:

Address Table

We need to partition the table. That means we need to split the data according to some logic. DANSE_RANK() function does it for us. It split users by their ID (UserID), order the partitions descending by date and adds ranks.

SELECT
	A.ID
	, A.UserID
	, U.FirstName
	, U.LastName
	, A.StreetNo
	, A.StreetName
	, A.City
	, A.Phone
	, A.DateOfRegistration
	, DENSE_RANK() OVER (PARTITION BY UserID ORDER BY DateOfRegistration DESC) AS [Rank]

FROM
	[Address] AS A
	INNER JOIN dbo.[Users] AS U
		ON A.UserID = U.ID

--WHERE
--	[Rank] = 2
GO

The result of the query is the data plus a temporary column for ranking.

Adding DENSE_RANK column (click on image)

If I try to reach the result in single query (uncommenting lines 18 and 19), I’ll get “Invalid column name ‘Rank’” error. The reason is that I generate the [Rank] column during the selection of the data and I can’t filter it at this time.

Invalid column name 'Rank'

To get the data that I need, I simulate a View – create a temporary table and read a filtered data from it.

USE zlika
GO

CREATE TABLE #Temp
(
	ID INT
	, [UserID] INT
	, [FirstName] VARCHAR(50)
	, [LastName] VARCHAR(50)
	, [StreetNo] INT
	, [StreetName] VARCHAR(100)
	, [City] VARCHAR(100)
	, [Phone] VARCHAR(25)
	, [DateOfRegistration] DATETIME NULL
	, [Rank] INT
)
GO

INSERT INTO #Temp
SELECT
	A.ID
	, A.UserID
	, U.FirstName
	, U.LastName
	, A.StreetNo
	, A.StreetName
	, A.City
	, A.Phone
	, DateOfRegistration, DENSE_RANK() OVER (PARTITION BY UserID ORDER BY DateOfRegistration DESC) AS [Rank]

FROM
	[Address] AS A
	INNER JOIN dbo.[Users] AS U
		ON A.UserID = U.ID
GO

SELECT
	ID
	, UserID
	, FirstName
	, LastName
	, StreetNo
	, StreetName
	, City
	, Phone
	, DateOfRegistration

FROM
	#Temp

WHERE
	[Rank] = 1

ORDER BY
	UserID
GO

DROP TABLE #Temp
GO

The result is the last address of the residents. If we need the precedent addresses, we’ll filter the [Rank] column to equals 2.

DENSE_RANK result

Félicitations

Feed enhanced by Better Feed from Ozh

Simple SSIS Project – export to flat file, FTP upload, archive (create the package)

zlika — Sun, 06 Jun 2010 18:26:29 +0000

Microsoft® SQL Server™ Integration Services is the tool that connects the database to the world and works instead of us. I’ll show you how to extract simple data from the database, store the files in a the file system, rename them, upload them on FTP and report the errors to administrator if any.

In this first part I will create the package. Next I’ll implement it into SQL Server.

The steps we need are:
1. Create the directories for storing the files – C:\Temp\Source\, C:\Temp\ForUpload\, C:\Temp\Archive\. If they already exist, we use them instead of create them
2. Move the files from Source directory to ForUpload directory with changing of the name (adding date and time)
3. Upload the renamed files on FTP server
4. Move the files to Archive folder

First, create a new SSIS project (File –> New –> Project) in SQL Server Business Intelligence Development Studio.

New SSIS Project

Rename the package.

Rename package

Create Data Source – that’s the place where we will read the data from – MS SQL Server. In Solution Explorer, right click on Data Sources –> New Data Source…

Create Data Source

To use the Data Source, I create a New Connection from the existing Data Source. Right click on Connection Managers –> New Connection From Data Source…

Create New Connection from Data Source

Create the global variables for the directories.

Create global variables

Create the directories – Source, ForUpload, Archive. Drag the File System Task from Toolbox pane.

Create Folders

Right click on File System Task to set the properties of the folders.

Create Folder

For Exporting the data to flat file, first, I create Data Flow Task.

Create Data Flow Task

Data Flow Tab –> Toolbox –> Data Flow Sources (ADO NET Source), Data Flow Transformation Source (Sort), Data Flow Destinations.

Edit ADO NET Source. I have written the query in SSMS, but I can build it on this step.

Edit ADO NET Source

Edit Sort

Edit Sort Transformation

Create Flat File Connection – Delimited

Create Flat File Connection

I chose the delimiter ($) in the Columns section

The next step is to rename the files by adding to filename the date and time. I use Foreach Loop Container to cover all the files in the Source folder. I move them to ForUpload folder and rename them in one step. I do it like this, because I need to know on which step the execution is failed and in that case to handle the rest of operations manually.

Add Foreach Loop Container

Create the local variables for Foreach Loop Container. We need to combine the file path with file name while looping the folder. In the Filename variable we will store the names of the files. To the variable Dir_ForUpload_Filename we will add the date and time.

Create local variables for File Loop Container

I create expressions for Dir_ForUpload_FileName and Dir_Source_FileName variables. When I select the variable, the Properties windows (in the down-right corner of the screen) changes.

Create expression for variable

In the Expression field I write the expresisons:

1. Dir_ForUpload_FileName:

@[User::Dir_ForUpload]
+ SUBSTRING(@[User::FileName], 1, FINDSTRING(@[User::FileName], ".", 1) - 1)
+ "_"
+ SUBSTRING((DT_WSTR, 30)GETDATE(), 1, 4)
+ "-"
+ SUBSTRING((DT_WSTR, 30)GETDATE(), 6, 2)
+ "-"
+ SUBSTRING((DT_WSTR ,30)GETDATE(), 9, 2)
+ "_"
+ SUBSTRING((DT_WSTR, 30)GETDATE(), 12, 2)
+ SUBSTRING((DT_WSTR, 30)GETDATE(), 15, 2)
+ SUBSTRING((DT_WSTR, 30)GETDATE(), 18, 2)
+ SUBSTRING(@[User::FileName], FINDSTRING(@[User::FileName], ".", 1), LEN(@[User::FileName]))

2. Dir_Source_FileName:

@[User::Dir_Source] +  @[User::FileName]

Edit Foreach Loop Container. We mark a folder for loop (C:\Temp\Source\). We loop only the .txt files. In Variable Mappings section we add the FileName variable.

Setting Foreach Loop Container

Setting File System Task (inside the Foreach Loop Container)

Setting File System Task

FTP upload – drag the Foreach Loop Container and FTP task in it. I create local variable FileName – the same as in previous step. In Foreach Loop Editor I choose Foreach File Enumerator, C:\Temp\ForUpload\ in Folder field and *.txt in Files field.

Edit FTP task – I adjust the local and remote paths

Setting FTP task

The next step is loop for placing the files in Archive folder. I won’t explain it. It’s the same as the loop, I used to move and rename from Source to ForUpload folder.

The full schema of the objects looks like this:

All the objects

Next I will implement the package into SSMS, I will create Scheduled Job, which will include operator notification. While executing the Job in real-time working enviroment, I will decide whether to add the Send Mail Task in SSIS package.

Source files

Félicitations

Feed enhanced by Better Feed from Ozh

The difference between WHERE and HAVING filters in SQL Server

zlika — Tue, 27 Apr 2010 02:44:51 +0000

The difference between WHERE and HAVING filters in Transact-SQL queries is that WHERE clause is a filter of all the data, while HAVING filters the GROUP BY data. I’ve prepared a simple example.

USE zlika
GO

-- Create table
CREATE TABLE test (
	Date DATETIME
	, QTY INT
)
GO

-- Populate table
INSERT INTO test (Date, QTY)
SELECT '2009-05-10', 1
UNION ALL SELECT '2009-05-10', 2
UNION ALL SELECT '2009-05-10', 3
UNION ALL SELECT '2009-05-11', 4
UNION ALL SELECT '2009-05-11', 5
UNION ALL SELECT '2009-05-11', 6
UNION ALL SELECT '2009-05-12', 7
UNION ALL SELECT '2009-05-12', 8
UNION ALL SELECT '2009-05-12', 9
GO

-- Check the table
SELECT
	CONVERT(CHAR(10), Date, 103) AS Date
	, QTY

FROM
	test
GO

The content of the table is as follows:

Content of the table

If we GROUP BY column QTY, the result is different sums of the QTY column:
May 10, 2009, 1 + 2 + 3 = 6
May 11, 2009, 4 + 5 + 6 = 15
May 12, 2009, 7 + 8 + 9 = 24

USE zlika
GO

-- GROUP BY (no filters)
SELECT
	CONVERT(CHAR(10), Date, 103) AS Date
	, SUM(QTY) AS QTY

FROM
	test

GROUP BY
	Date
GO

GROPU BY filter

If we exclude the ’6′ in the WHERE clause, it’s gonna be filtered before the grouping of the date. The line No. 6 will be eliminated and the May 11, 2009 will be grouped like 4 + 5 = 9.

USE zlika
GO

-- GROUP BY (WHERE filter)
SELECT
	CONVERT(CHAR(10), Date, 103) AS Date
	, SUM(QTY) AS QTY

FROM
	test

WHERE
	QTY <> 6

GROUP BY
	Date
GO

WHERE filter

If we exclude the ’6′ in the HAVING clause, it’s gonna be filtered after the grouping of the date. The date that has sum of QTY column = 6 will be eliminated and the date May 10, 2009 will miss.

USE zlika
GO

-- GROUP BY (HAVING filter)
SELECT
	CONVERT(CHAR(10), Date, 103) AS Date
	, SUM(QTY) AS QTY

FROM
	test

GROUP BY
	Date

HAVING
	SUM(QTY) <> 6
GO

HAVING filter

Clean up the test table.

USE zlika
GO

-- Clean up
DROP TABLE test
GO

Félicitations

Feed enhanced by Better Feed from Ozh

MS SQL Server Cursor Example

zlika — Sat, 06 Mar 2010 05:35:09 +0000

The cursors in MS SQL Server are not the best way to manipulate the data, because they need more recourse. If you can escape the usage of cursors, do it. OK, the cursors give us a flexible way of manipulating the information. You can imagine a bundle of information, that we use for move through other information and modify it using the information of our bundle (puting the information in the WHERE clause).

Let’s say we need to modify a table containing user information. We need to mark those of them, who are registered more than a month (or 30 days) ago. Simultaneously, we need to write in another table only the modified users and the time of modification.

First we create tables, add keys and populate tables.

USE zlika
GO

-- Create Tables
CREATE TABLE Departments
(
	ID INT NOT NULL
	, DepartmentName VARCHAR(50)
)
GO

CREATE TABLE Person
(
	ID INT IDENTITY
	, FirstName VARCHAR(20)
	, LastName VARCHAR(50)
	, DepartmentID INT
	, RegisteredDate DATETIME
	, CheckStatus BIT
)
GO

CREATE TABLE Checked
(
	PersonID INT
	, CheckDate DATETIME
)
GO

-- Add Keys
ALTER TABLE Departments
ADD CONSTRAINT PK_DepartmentID PRIMARY KEY CLUSTERED
(
	ID ASC
)
GO

ALTER TABLE Person
WITH CHECK ADD CONSTRAINT
FK_Departments_DepartmentID FOREIGN KEY
(
	DepartmentID
)
REFERENCES Departments
(
	ID
)
GO

-- Populate Tables
INSERT INTO Departments (ID, DepartmentName) VALUES
(1, 'IT')
, (2, 'Administration')
, (3, 'Technical Support')
, (4, 'Accounting')
, (5, 'Marketing')
GO

INSERT INTO Person (Firstname, LastName, DepartmentID, RegisteredDate) VALUES
('First', 'Smith', 5, '2010-02-01')
, ('Second', 'Brown', 4, '2010-02-02')
, ('Third', 'Marchal', 3, '2010-02-03')
, ('Fourth', 'Jameson', 2, '2010-02-04')
, ('Fifth', 'Anderson', 1, '2010-02-05')
, ('Sixth', 'Cameron', 1, '2010-02-06')
, ('Seventh', 'Blake', 2, '2010-02-07')
, ('Eight', 'Robert', 3, '2010-02-08')
, ('Ninth', 'Jameson', 4, '2010-02-09')
, ('Tenth', 'Wiliams', 5, '2010-02-10')
GO

-- Preview Tables
SELECT
	*

FROM
	dbo.Person AS P INNER JOIN dbo.Departments AS D
		ON P.DepartmentID = D.ID
GO

Create and populate tables

Let’s say we need to mark only the users of IT, Technical Support and Marketing departments who are registered 28 days ago.

USE zlika
GO

SELECT
	P.FirstName
	, P.LastName
	, D.DepartmentName
	, CONVERT(CHAR(10), P.RegisteredDate, 103) AS RegisteredDate

FROM
	dbo.Person AS P INNER JOIN dbo.Departments AS D
		ON P.DepartmentID = D.ID
GO

Select all persons

The information that we’re going to use as filter (in the WHERE clause) is PersonID, RegisteredDate from Person table. That’s the query that is going to populate the cursor.

USE zlika
GO

SELECT
	P.ID
	, P.RegisteredDate

FROM
	dbo.Person AS P INNER JOIN dbo.Departments AS D
		ON P.DepartmentID = D.ID

WHERE
	D.DepartmentName IN ('IT', 'Technical Support', 'Marketing')
	AND P.RegisteredDate < DATEDIFF(DAY, 28, GETDATE()) -- 2010-02-06
GO

Populate Cursor Statement

The work we need to do is between BEGIN and END keywords. First we mark the status as ‘checked’ (1), then we add the PersonID and current date in another table (Checked).

USE zlika
GO

-- Execute the UPDATE Statement
UPDATE Person
SET CheckStatus = 1
WHERE ID IN (1, 3, 5)

-- Execute the INSERT Statement
INSERT INTO Checked (PersonID, CheckDate) VALUES
(1, GETDATE())
, (3, GETDATE())
, (5, GETDATE())

OK, here’s the guy who do the real work – the cursor itself. We store the content of the cursor in a variables and we execute the two steps for every row of it. The variables are our filters – we use them in the WHERE clause.

USE zlika
GO

-- Declare the variables
DECLARE @PersonID INT
DECLARE @Date DATETIME = DATEDIFF(DAY, 28, GETDATE())

-- Declare the cursor
DECLARE Update_Person_Cursor CURSOR FOR

-- Populate the cursor
SELECT
	P.ID
	, P.RegisteredDate

FROM
	dbo.Person AS P INNER JOIN dbo.Departments AS D
		ON P.DepartmentID = D.ID

WHERE
	D.DepartmentName IN ('IT', 'Technical Support', 'Marketing')
	AND P.RegisteredDate < @Date
	AND P.CheckStatus IS NULL

-- Open the cursor to work with it
OPEN Update_Person_Cursor

-- Select The first row into the variables (@PesronID = 1, @Date = 2010-02-01)
FETCH NEXT FROM Update_Person_Cursor INTO
	@PersonID
	, @Date

-- While there are no more rows in the cursor...
WHILE @@FETCH_STATUS = 0

-- Do the two steps!
BEGIN
	-- Execute the UPDATE statement
	UPDATE Person
	SET CheckStatus = 1
	WHERE ID = @PersonID

	-- Execute the INSERT statement
	INSERT INTO Checked (PersonID, CheckDate) VALUES
	(@PersonID, GETDATE())

	-- Select The second row into the variables (@PesronID = 3, @Date = 2010-02-03)
	FETCH NEXT FROM Update_Person_Cursor INTO
		@PersonID
		, @Date
END

-- Close The cursor
CLOSE Update_Person_Cursor

-- Kill the cursor
DEALLOCATE Update_Person_Cursor
GO

The records in the ‘Checked’ table are those we needed.

USE zlika
GO

SELECT * FROM Checked

Content of 'Checked' table

…and the status in the ‘Person’ table is changed to ’1′.

USE zlika
GO

SELECT
	*

FROM
	dbo.Person AS P INNER JOIN dbo.Departments AS D
		ON P.DepartmentID = D.ID
GO

The content of 'Person' table

Félicitations

Feed enhanced by Better Feed from Ozh

Generate filename (Name_Date_Time.txt) for a report with Transact-SQL in Microsoft SQL Server

zlika — Mon, 01 Mar 2010 03:18:58 +0000

My ex-colleague asked me how to insert query result in a variable for generate report in Microsoft SQL Server. After few explanations I noticed that she has decided to use SSIS (SQL Server Integration Services) to automate. She sent me an example in which she uses YEAR(), MONTH() and DAY() functions.

-- YEAR(), MONTH(), DAY()
SELECT
	'Test_'
	+ CAST(YEAR(GETDATE()) AS CHAR(4))
	+ '-'
	+ CAST(MONTH(GETDATE()) AS CHAR(2))
	+ '-'
	+ CAST(DAY(GETDATE()) AS CHAR(2))
	+ '.txt' AS DynamicFileName
GO

Result

The result is strange, because the return data type of those functions is integer. To fix this, we need to add leading zeros in front of days, months, hours, minutes and seconds.

-- Add leading zeroes
DECLARE
	@Day AS CHAR(2)
	, @Month AS CHAR(2)
	, @Year AS CHAR(4)
	, @Hour AS CHAR(2)
	, @Minute AS CHAR(2)
	, @Second AS CHAR (2)
	, @NewFileName AS CHAR(26)

SELECT @Day = DATEPART(DD, GETDATE())
SELECT @Day = RIGHT('0', 2 - LEN(@Day)) + @Day

SELECT @Month = DATEPART(MM, GETDATE())
SELECT @Month = RIGHT('0', 2 - LEN(@Month)) + @Month

SELECT @Year = DATEPART (YY, GETDATE())

SELECT @Hour = DATEPART(HH, GETDATE())
SELECT @Hour = RIGHT('0', 2 - LEN (@Hour)) + @Hour

SELECT @Minute = DATENAME(MI, GETDATE())
SELECT @Minute = RIGHT('0', 2 - LEN (@Minute)) + @Minute

SELECT @Second = DATEPART(SS, GETDATE())
SELECT @Second = RIGHT('0', 2 - LEN (@Second)) + @Second

SELECT @NewFileName = 'Test_' + @Year + '-' + @Month + '-' + @Day + '_' + @Hour + @Minute + @Second + '.txt'

SELECT @NewFileName AS DynamicFileName
GO

Result

The best way is to convert the result of GETDATE() function to CHAR and extract the different parts of it with the SUBSTRING() function.

-- SUBSTRING()
DECLARE
	@Date CHAR(20) = CONVERT(CHAR(10), GETDATE(), 112) -- yyyymmdd
	, @Time CHAR(20) = CONVERT(CHAR(10), GETDATE(), 108) -- hh:mi:ss

SELECT
	'Test_'
	+ SUBSTRING(@Date, 1, 4) -- Year
	+ '-'
	+ SUBSTRING(@Date, 5, 2) -- Month
	+ '-'
	+ SUBSTRING(@Date, 7, 2) -- Day
	+ '_'
	+ SUBSTRING(@Time, 1, 2) -- Hour
	+ SUBSTRING(@Time, 4, 2) -- Minute
	+ SUBSTRING(@Time, 7, 2) -- Second
	+ '.txt' AS DynamicFileName
GO

Result

My ex-colleague needs the dynamic filename to generate reports with SSIS. I’m gonna show her soon how to create simple project and the way for the same manipulation of the date in SSIS.

Félicitations

Feed enhanced by Better Feed from Ozh