Introduction to sql server

Introduction to
SQL Server
(For Fresher)
As the title indicates this is only for fresher who has just started the carrier or who is
waiting to open up the carrier in the application programming side. But that does not
mean this article is a tutorial for a fresher who does not know anything about SQL.
This article is meant for who already have a little knowledge in SQL and want to
improve it.
The prerequisite for learning SQL is knowledge in Discrete Mathematics (Set Theory,
Relations and Functions). Although it is not necessary to learn all the theorems and
proof for the theorems in the Discrete Mathematics, you should have learned the
basic concepts of the Sets, Relations and Functions. This will help you to learn SQL
queries and fundamentals easily. If you want to explore a RDBMS more deeply you
should learn Graph Theory too.
Although I tried to avoid SQL Server specific topics in this article, I am sure that
some topics are pure to SQL server such as SQL Enterprise manager.
By Samuel

http://www.samueldotj.com/Articles.asp
Data to DBMS
Data is something that should be stored for future manipulations (In terms of
Database). The system which provides such a facility is called Database Management
System or DBMS.
The simplest form to store a data for latter retrieval is using a text file. For example
you may want to store your friends name and phone numbers to use it latter. (In
this case you may use notepad or word to do so.) This kind of storage is called flat
file storage or unstructured storage. In this case the text editor uses the File and
Directory services provided by the Operating System to accomplish the task of
storing and retrieving data.
But these unstructured flat files are not suitable to large data such as storing stock
details. Since the stock data is large in volume and added and updated frequently it
is not scale up well if we use a simple flat file. To overcome this we need some
system which should perform the storing, retrieving, manipulating and querying
operations on the data and give output to us. This kind of system is called Database
Management System. So a DBMS is a system which organizes and stores data in a
structural way for fast retrieval. A DBMS uses one or more files to store the given
data.
Since the DBMS is meant mainly for developers every DBMS has it is own language
to write the commands. The languages are standardized under the name SQL
(Structured Query Language). The following definition for SQL is from Books Online
“A language used to insert, retrieve, modify, and delete data in a relational database.
SQL also contains statements for defining and administering the objects in a
database. SQL is the language supported by most relational databases, and is the
subject of standards published by the International Standards Organization (ISO)
and the American National Standards Institute (ANSI). SQL Server 2000 uses a
version of the SQL language called Transact-SQL.”

Representation of Data by DBMS
DBMS should represent the data stored by them in some form and it is most
common to represent them as Column, Row, Tables and Databases. As you know
SQL you know what are columns, rows, tables and database. However I want to give
one line explanation for those things here.
Column – A finite unit of data. It is represented by one of the data type.
Row – A collection of Columns.
Tables – Collection of rows.
Database – Collection of tables and other objects.
For example in case of Employee records – the name, date of join, salary etc will be
called as Column or Field. The value for a particular employee for the above field is
called as Row or Record.
Primary Key
A column or set of columns used to uniquely identify the records in a table. Primary
Keys don’t allow NULL values. You can relate a table with other table if you have
defined primary on it. (So defining UNIQUE and NOT NULL constraint is not
equivalent to primary key). SQL server will create a Clustered index while you create
a primary key.
Primary keys should be carefully defined since based on it other relations are
defined. Poorly designed primary keys will affect insert, update and delete
operations. Primary keys are different from which are used for paperwork. For
example – you can have unique EmployeeNo(varchar) field. It may be of the format
XXXX##XXX##. Although it is unique, making it primary key is not a good design
because it is of character type and it may be clustered. Instead you should use it as
alternate key and create primary key EmployeeId (int).
Since primary keys needs to be unique, SQL server provides two methods to
generate unique keys – IDENTITY and UNIQUEIDENTIFIER.

Server - Client
In Client Server technology, a server is a program which does all the jobs that is
requested by the client and Client is a program which depends on the Server to
accomplish its task. Client is mainly used to interact with Users and most Server
software doesn’t have direct interaction with Users.
A DBMS should have the following functionality – Interaction with users (developers),
Managing Data. Since these two functionalities are entirely different modern DBMS is
divided into two or more parts. Almost all current DBMS has the following module –
Server and one or more clients. A server module is program which manages the data
and handles requests from the client applications. It interprets the SQL queries and
returns results back to the clients. A client program is a program which sends SQL
commands to server and gets the result. Client program does not know the detailed
underground structure of the server.
SQL Server 2000 runs as a service in NT based machines. You have to establish a
connection to the server first if you want to do some SQL operation. SQL Server uses
Local or Network IPC (Inter process communication) to communicate with the
clients. See Books Online for more details.
The most basic misunderstanding of many junior developers is considering SQL
Enterprise Manager as the SQL server. SQL Enterprise Manager is a tool used to
access the SQL server. SQL Query analyzer is also similar tool used to access the
SQL server. The difference between SQL query analyzer and Enterprise Manager is
Query Analyzer is a light weight process and Enterprise manager is heavy weight
processes due to its Graphical User Interface.
Enterprise Manager is a handy tool in the initial stage of a project. It will save lot
more time while designing the database. You can easily create Tables, Triggers,
Relationships, Constraints etc using Enterprise manger easily. I have seen many of
developers don’t have experience in using Enterprise Manager. Never become one of
them, use Enterprise Manager whenever you start creating a database. Query
Analyzer is a tool used to Create, Analyze, Modify and Delete T-SQL queries.

You can also access SQL server using osql.exe or isql.exe which is a very light weight
command line utility. You can execute T-SQL commands in osql/isql. The difference
between osql and isql is osql uses ODBC library whereas isql uses DB library to
communicate with the SQL server. SQL Query analyzer is Windows version of
isql.exe.
You can directly access SQL server by accessing the TCP Port/Named Pipes and
issuing certain commands. However it is very tedious, so SQL Server has provided
some library to access it easily. You can find the source code for these libraries in the
installation directory of the SQL Server. For database communication standards SQL
Server fully supports ODBC. Further it has own database driver for communication.
SQL Server also offers SQL-DMO (SQL Distributed Management Objects) a COM
component which offers easy programming interface. I have created one application
called Whiz using the SQL-DMO. You can download source code for that application in
my projects page.
Apart from these libraries many third party libraries are also available to access SQL
Server. In .Net you can directly access SQL server using Data.SQLClient namespace.
Now you have learned how SQL Server and SQL Client tools are exists in a network.
It is time to learn some Client-Server technology rules
1) There may be more than one connection to server at any time
2) Client should request only what it wants
3) Client requests should be short
The first point is related to Server so leave it. The second point says that the client
should retrieve only the data that it needs. This is because transferring data from
Server to Client utilizes server’s resource and also increases network traffic.
Here is an example, you have an employee table in the server and you want to see
Employee’s name who has joined before ‘1-1-1990’. You can do it using the following
three queries. The last query is best for Client-Server topology.
SELECT * FROM Employee
SELECT * FROM Employee WHERE DOJ < ‘1-1-1990’
SELECT EmpName FROM Employee WHERE DOJ < ‘1-1-1990’

DDL (Data Definition Language)
Data Types
What is data type? – Classification of data into similar groups. Names, Money,
Date, Time, etc are examples for data type.
What is the use of classification or data type? – It increases the performance,
reduces the space needed to store the data.
I can’t explain how it increases the performance and how it reduces the space in a
single paragraph. However I can direct you to find the solution – Try writing a C
program which should store Employee Name, DateOfBirth, Salary, EmployeeID into
file and provide option to search the records.
The following data types are available in SQL.
TINYINT, SMALLINT, INT, BIGINT
DECIMAL/NUMERIC
FLOAT, REAL
SMALLMONEY, MONEY
BIT
CHAR, NCHAR, VARCHAR, NVARCHAR, TEXT, NTEXT
BINARY, VARBINARY, IMAGE
CURSOR
SQL_VARIANT
TABLE
TIMESTAMP
UNIQUEIDENTIFIER
The main difference between the data type is the capacity of the type. To see the
capacity of the data type refer “Books Online”.
CREATE TABLE, Constraints, Relationships,

All your definitions of the tables, views, stored procedures, functions are stored in
system tables. Books online - The information used by Microsoft® SQL Server™
2000 and its components is stored in special tables known as system tables.
You can query the system tables to obtain the inside structure of your schema. The
following system tables are the some of the most used system tables – sysobjects,
syscolumns, systypes, sysusers, sysindexes. Some of the tables are stored only in
master database because they contain contents which are global across the
databases – syslogins, syslockinfo, sysdatabases etc.
For example to get list of databases in the connected server, you can issue –
SELECT * FROM sysdatabases
The following query will return all the table names in the connected database.
SELECT TableName = [Name] FROM sysobjects WHERE xtype=’U’
The following query will return all the column details for a given Table.
SELECT *
FROM sysobjects so(NOLOCK)
INNER JOIN syscolumns sc(NOLOCK)
ON so.[id]=sc.[id]
WHERE so.[Name]=’TableName’
You can query the system tables but you should not add, modify or delete anything
directly. To add, delete and modify system table contents SQL Server provides some
stored procedures.
sp_databases – This stored procedure will return all the databases in the connected
server.
sp_tables – This will return all the tables in the database.
sp_columns – This stored procedure will return the column details of a given Table.

DML (Data Manipulation Language)
Data Manipulation Languages contains command used to query, change, and
add data from/to the database. This includes the following commands - SELECT,
INSERT, DELETE. All these commands are deals with set of data. So I want to refresh
the Set theory concepts before going further.
Set
A set is a collection of objects considered as a whole.
Subset
If every member of the set A is also a member of the set B, then A is said to
be a subset of B
Union of two sets
The union of A and B, denoted by A U B, is the set of all things which are
members of either A or B.
Intersection of two sets
The intersection of A and B, denoted by A ∩ B, is the set of all things which
are members of both A and B. If A ∩ B = ø, then A and B are said to be disjoint.
To read more about Set theory visit http://en.wikipedia.org/wiki/Set
In SQL server, all data is handled as sets – rowset or recordset. Hereafter, the word
rowset or recordset refers a collection of data.

SELECT
This selects/extracts/returns only one rowset from the datasource. The datasource
may be anything. In short, if you want to extract any data from SQL server you
should use SELECT. Similarly, if you want to return any data to user or client
application you should use SELECT. (Note - SELECT is also capable of inserting
records into a table or XML file)
SELECT * FROM Employee
The above command instructs the SQL Server to extract all the columns and all the
rows from the Employee table. But in practice you should not write a query like this –
because it extracts all the columns and all the rows which may return GBs of data.
Consider the employee table has the following fields Name, Address and Photo. Since
the photo column is a binary image the column size may be more than 100KB. If
there are one thousand employees and then the picture column alone will take
1000*100KB = 100MB and the remaining name and address fields will take less than
1MB. If you want to see only Employee Name and address then the above query will
return 101MB of data from which only 1MB is worth for you. So select only the fields
which are required.
SELECT Name, Address FROM Employee
This command instructs the SQL Server to extract only name and address columns
and all the rows from the Employee table. Again you should not use queries like this
without row constraint. You can add row constraint to a query in two ways – TOP and
WHERE clause.
TOP Clause
This query will return the first seven Employees from the database. Although it is not
very useful in any real time projects, it is very handy during debugging or data
analysis.
SELECT TOP 7 Name, Address FROM Employee
The TOP clause is very useful when you combine it with ORDER BY clause. For
example the following query will return the 7 senior most employees in the
organization.

SELECT TOP 7 Name, Address, DateOfJoin, BasicPay FROM Employee ORDER BY
DateOfJoin
As the query starts growing the complication starts because the clarity in reading the
query is lost. To minimize this some kind of coding standard should be followed. I am
following a standard which has the following rules
1) All SQL keywords should be in capital letter
2) Command should in the first line (with Top or any other clause)
3) Each column names should be in a separate line with one indent
4) The data source should be in a separate line
So the above query can be written as follows
SELECT TOP 7
Name,
Address,
DateOfJoin,
BasicPay
FROM Employee
ORDER BY DateOfJoin
SET ROWCOUNT option can also used to restrict the number of rows returned.

WHERE clause
To add one and/or more condition to a query the WHERE clause is used.
SELECT
Name,
Address,
DateOfJoin
FROM Employee
WHERE DateOfJoin > ‘1-Jan-2000’
The above query will return who has joint after 1-Jan-2000. The following query will
return who has joined after 1-Jan-2000 and before 12-Dec-2000.
SELECT
Name,
Address,
DateOfJoin,
BasicPay
FROM Employee
WHERE DateOfJoin > ‘1-Jan-2000’ AND DateOfJoin < ‘12-Dec-2000’
Or you can write it as
SELECT
Name,
Address,
DateOfJoin
FROM Employee
WHERE DateOfJoin BETWEEN ‘1-Jan-2000’ AND ‘12-Dec-2000’
Now try to write a query that should return the last employee details who has joined
before ’1-Jan-2000’
SELECT TOP 1
Name,
Address,
DateOfJoin
FROM Employee
WHERE DateOfJoin > ‘1-Jan-2000’
ORDER BY DateOfJoin DESC

GROUP BY
It is a clause which is used to summarize or group the data. For example, if you look
into a cricket score board you may want to know the total runs instead of individual
batsman scores.
SELECT
Department,
TotalPay = SUM(BasicPay)
FROM Employee
GROUP BY Department
Here I have summarized the salary amount paid for each department.
SELECT
Department,
TotalEmployees = COUNT(*)
FROM Employee
GROUP BY Department
The above query will return total employees in a department.
Now try to write a query that should give DepartmentName and average salary
without using the AVG() function.
INTO
The INTO keyword is used to create a table and insert the selected records into the
newly created table.
SELECT
Department,
TotalEmployees = COUNT(*)
INTO #Result
FROM Employee
GROUP BY Department

Multiple Rowsets
Up to now I have shown you how to deal with a single result set. Now let try to add
or substract one or more rowsets to create new rowsets. For this assume you have 3
tables with employee information – Employee_India, Employee_USA,
Employee_China.
Now write query to select all the employee name from the 3 employee tables.
SELECT Name FROM Employee_India
UNION
SELECT Name FROM Employee_USA
UNION
SELECT Name FROM Employee_China
The above query will return all the name from the three employee tables. Now take
only the names which starts with the letter ‘A’.
You can do it in two ways adding WHERE clause to all the three tables or adding one
WHERE clause to the final rowset
SELECT Name FROM Employee_India WHERE Name LIKE ‘A%’
UNION
SELECT Name FROM Employee_USA WHERE Name LIKE ‘A%’
UNION
SELECT Name FROM Employee_China WHERE Name LIKE ‘A%’
Or
SELECT Name
FROM
(
SELECT Name FROM Employee_India
UNION
SELECT Name FROM Employee_USA
UNION
SELECT Name FROM Employee_China
)
WHERE Name LIKE ‘A%’

JOINS
Relations between two row set is referred as JOIN in SQL. JOIN can be any one of
the following types – one to one, one to many, many to many. Learn more about
these types of relations in Discrete Mathematics. Similarly a function may be
surjective, injective or bijective.
When I was start learning T-SQL I was ignoring the JOIN keyword syntax used by
SQL Server, instead I was used to join tables using WHERE condition. I have seen
many junior developers are doing this way. This is because the JOIN syntax is seem
to be slightly confusing (at the starting stage) and it is only specific to MS-SQL
Server. But once you understood the JOIN keyword syntax it is much very much
useful to maintain the query.
To explain joins assume the following table structure.
If you are not able to understand this table structure
then you have learn lot of fundamentals before
continuing.
Here the relation between the Student and Mark table is Student_id, similarly the
relation between Subject and Mark table is Subject_id.
You can extract marks obtained by student using the following query.
SELECT s.Name, m.Mark
FROM Student s, Mark m
WHERE s.Student_id = m.Student_id
The following query will return the same result set but using the JOIN syntax which
is SQL-92 standard
FROM Student s INNER JOIN Mark m
ON s.Student_id = m.Student_id
Student
Student_id Int
Name Varchar(20)
Subject
Subject_id Int
Name Varchar(20)
Mark
Student_id Int
Subject_id Int
Mark Tinyint

The difference between relationship and condition is very clear in JOIN syntax where
it is not clear in the ordinary query. For example you have to retrieve Student Names
who has scored more than 50 then you can write the following queries.
FROM Student s, Mark m
WHERE
s.Student_id = m.Student_id
AND m.Mark > 50
In this query the relationship (s.Student_id = m.Student_id) and condition is
(m.Mark > 50) is specified in WHERE condition which will make the query
complicated. The following query is replacement of the above query.
FROM Student s INNER JOIN Mark m
WHERE
m.Mark > 50

OUTER JOIN
Some times we may found that the relation between two set is surjective. That is
there may be elements in either or both side of the sets that is not linked with other
set. For example consider the following example –
Emp_Id Emp_Name Dept_Id
1 A 1
2 B 1
3 C 2
4 D NULL
5 E 1
Dept_Id Dept_Name
1 X
2 Y
3 Z
In the employee table the Employee D (Emp Id - 4) does not related to the
Department table. Similarly the department Z is not related the Employee table.
Now try to write to a query that should return the following fields – Employee Name,
Department Name. The following query will miss out the Employee D.
SELECT
EmployeeName,
DeptName
FROM Employee E
INNER JOIN Department D
ON E.Emp_id = D. Emp_id
To include the all the records in the left table you have to use LEFT OUTTER JOIN,
similarly you have to use RIGHT OUTTER JOIN to include all the values in the right
table.
SELECT
EmployeeName,
DeptName
FROM Employee E
LEFT OUTTER JOIN Department D

In this query instructs to include all the values from the Employee table and then link
it to department table. LEFT OUTTER JOIN works based on the position of the table
name that is the table placed left to the OUTER JOIN key word will include in the
whole. Sometimes when you join more than 2 tables you may need to use RIGHT
OUTTER JOIN because the table may come RIGHT to your query.
Now try to write a query which should include all employee records and department
names even if they are not related.
The following query will display all the employee name and department name.
SELECT
EmployeeName,
DeptName
FROM Employee E
FULL OUTTER JOIN Department D
The result will be
EmployeeName DeptName
A X
B X
C Y
D NULL
E X
NULL Z

CROSS JOIN
Cross join will produce Cartesian product from two set. That is it will generate all the
possible combination of relation between two set. The difference between CROSS
JOIN and FULL OUTTER JOIN is full outer join tries to link the elements and include
the elements, but cross join links each element in one set with all other element in
the other set. CROSS JOIN is equivalent to linking two tables without any join
condition.
The following two queries will produce same result
SELECT EmpName, DeptName FROM Employee, Department
SELECT EmpName, DeptName FROM Employee CROSS JOIN Department
The result will be
EmployeeName DeptName
A X
A Y
A Z
B X
B Y
B Z
C X
C Y
C Z
D X
D Y
D Z
E X
E Y
E Z

UPDATE
UPDATE statement is used to modify one set of records at a time. For example you
can add title to the name of employees using the following query.
UPDATE Employee SET EmpName = ‘Mr.’ + EmpName
But it is not very useful until you add a condition. You can add a condition by using
the WHERE condition.
UPDATE Employee SET EmpName = ‘Mr.’ + EmpName WHERE Sex=’M’
UPDATE Employee SET EmpName = ‘Mrs.’ + EmpName WHERE Sex=’F’
The above query will update the Employee name with the prefixed title. The title will
be Mr. for all male employees and Mrs. for all female employees. (I assumed the
employer has a policy of recruiting only married persons). You can write the above
two queries in a single statement using CASE keyword.
UPDATE Employee
SET EmpName =
CASE Sex
WHEN ‘M’ THEN ‘Mr.’
WHEN ‘F’ THEN ‘Mrs.’
END
+ EmpName
Note you can CASE statement in any SQL statement not only in UPDATE statements.
Another important feature of UPDATE statement is you need not use any temporary
variable to swap value between two fields. For example, assume in your employee
table the DateOfJoin field contains DateOfBirth data and DateOfBirth field contains
DateOfJoin field data. Now you have to swap the data, the following query will do
that –
UPDATE Employee
SET
DateOfJoin = DateOfBirth,
DateOfBirth = DateOfJoin
In very rare occasion you will need to write simple update queries like this, the more
frequent occasion will be combining two or more tables and updating one table.
Remember you can’t update more than one table at time. SQL server simply doesn’t
support. You can update only one table at a time. Remember the Student Table

structure that I have mentioned, now add another field called “Percentage (float)” to
the Student table. The following query will update it.
UPDATE s
SET Percentage = SUM(m.Marks) / COUNT(m.Marks)
FROM Student s
INNER JOIN Mark m
GROUP BY s.Student_id
In the above query you are updating the student table since you have mentioned the
alias name.
DELETE
Delete is easiest statement in the SQL, I think. To delete all records from a table you
can issue DELETE FROM TableName. Here we have not used DELETE * FROM
TableName because the * is used to specify all columns, since we can delete only
rows and not columns. The DELETE operation creates logs and can be undoable.
Since it performs under transaction it is safer but slower. The alternative command
to delete records from a table is TRUNCATE TABLE TableName. The TRUNCATE
command deletes all records from the table you cannot control it. In DELETE
command you control the delete operation by specifying the WHERE condition.

INSERT
Insert command is used to insert data into table.
The following commands will do the same thing. It will insert a new student name
into the table with student id 10001.
INSERT INTO Student (Student_id, StudentName)
VALUES (10001, ‘Vimal’)
If you are going to insert all values you can omit the field names.
INSERT INTO Student
VALUES (10001, ‘Vimal’)
(If you try to execute the same query second time you should get Primary Key
violation, if you have set primary key on the student_id field)
Assume you have table called OldStudent which has old student names and you want
to transfer it to the student table, the following query will do it.
SELECT Student_id, StudentName FROM OldStudent
Similarly assume you have student information in the following tables – ClassA,
ClassB, ClassC and you want to import into a Student Table you can use the
following query. (Add a field Class (Varchar) to student table, if you want to store
Class name also.)
INSERT INTO Student (Student_id, StudentName, Class)
(
SELECT Student_id, StudentName, ‘A’ FROM ClassA
UNION
SELECT Student_id, StudentName, ‘B’ FROM ClassB
UNION
SELECT Student_id, StudentName, ‘C’ FROM ClassC
)
The above queries will work fine only if the student_ids are unique across the table.
If the student_ids are auto-generated you can’t use them to insert into the Student
table because they will generate duplicate values and the query will fail. To avoid
that you have to generate Student_ids automatically, the following query will do
that.

SELECT
IDENTITY(int, 10000) AS Student_id,
StudentName
INTO #tmp
FROM OldStudent
SELECT Student_id, StudentName FROM #tmp
The identity function inserts serial numbers starting from 10000 into the Student
Table.
If you have inserted a single row into a table, then you may want to know the last
inserted identity value. For that SQL server provides the following one function and
one global variable - IDENT_CURRENT and @@IDENTITY
IDENT_CURRENT (TableName) returns the last identity value generated for a
specific table in any session and any scope.
@@IDENTITY returns the last identity value generated for any table in the current
session, across all scopes.

Transactions
A transaction is a sequence of operations performed as a single logical unit of work.
A logical unit of work must exhibit four properties, called the ACID (Atomicity,
Consistency, Isolation, and Durability) properties, to qualify as a transaction.
Atomicity and Consistency
SQL Server provides Data Control Language for controlling transactions. BEGIN
TRANSACTION is used to start a transaction. You can use ROLLBACK
TRANSACTION to rollback your transaction. ROLLBACK TRANSACTION erases all
data modifications made since the start of the transaction or to a savepoint. It also
frees resources held by the transaction. COMMIT TRANSACTION can be used to
commit your changes to the database. SQL server supports nested transactions.
BEGIN TRANSACTION
UPDATE Student SET StudentName =’Test’ WHERE StudentID = 10000
SELECT StudentName FROM Student WHERE StudentID = 10000
ROLLBACK TRANSACTION
The above query will return Test and OldStudentName, since the transaction is
cancelled.
BEGIN TRANSACTION
UPDATE Student SET StudentName =’Test’ WHERE StudentID = 10000
COMMIT TRANSACTION
The above query will return Test and Test since the transaction is updated.

Isolation
SQL Server provides locking hints to isolate a unit of work. A range of table-level
locking hints can be specified using the SELECT, INSERT, UPDATE, and DELETE
statements to direct Microsoft® SQL Server™ 2000 to the type of locks to be used.
NOLOCK, HOLDLOCK, ROWLOCK, TABLOCKX and UPDLOCK are the some of the
examples for lock hints.
1)
BEGIN TRANSCATION
UPDATE Student WITH (HOLDLOCK)
SET StudentName=’Test’
WHERE Student_id= 10000
SELECT StudentName FROM Student
WHERE Student_id = 10000
COMMIT TRANSACTION
2)
SELECT StudentName FROM Student WITH (NOLOCK)
WHERE Student_id = 10000
To test the isolation, run the 1st
query (comment out the COMMIT statement) in a
query analyzer, open another window of query analyzer and try to run the following
command – SELECT StudentName FROM Student WHERE Student_id = 10000. You
will be blocked until the transaction running in 1st
window is committed or roll
backed. If you want retrieve records without being blocked use the 2nd
query WITH
NOLOCK.
Durability
It is all about internal database logs. So it may not useful for a fresher.

Views
Abstraction and Security is the main purpose for which I have used Views.
Abstraction – You can hide the internal table structure from the client applications
and give a better interface through views. For example you can combine Student and
Marks table to get results and define it as a view as follows.
CREATE VIEW Result AS
SELECT
StudentName = s.StudentName,
TotalMark = SUM(m.Mark),
Grade = CASE
WHEN SUM(m.Mark)/COUNT(m.Mark) > ‘80’ THEN ‘First Class’
WHEN SUM(m.Mark)/COUNT(m.Mark) > ‘60’ THEN ‘Second
Class’
WHEN SUM(m.Mark)/COUNT(m.Mark) > ‘50’ THEN ‘Third Class’
ELSE ‘FAIL’
END
FROM Student s
INNER JOIN Mark m
Here the Relationship between tables and the calculations are abstracted.
Views are constructed as virtual table or logical table. This means that they does not
exists in the physical data page, they exists only in the code form. But when you run
a query on a view it is constructed as a virtual table in the TempDB.
You can create Clustered indexes on the Views.

Stored Procedures
A stored procedure is a group of Transact-SQL statements compiled into a single
execution plan. It is like a function or procedure in ordinary programming language.
You can return one more values from a stored procedure.
CREATE PROCEDURE sp_Test
AS
SELECT ‘Test’
SELECT StudentName, Student_id FROM Student
GO
The above stored procedure will return the following values – one ‘Test’ and all the
StudentName, Student_id from the student table.
CREATE PROCEDURE sp_StudentInsert
@StudentName AS VARCHAR(50), @StudentID AS INT OUTPUT
AS
IF LTRIM(@StudentName) = ‘’
BEGIN
RAISERROR(‘Student Name should be atleast one character’)
RETURN
END
IF EXISTS(SELECT 1 FROM Student WHERE StudentName = @StudentName)
BEGIN
RAISERROR(‘Student Name already exists’)
RETURN
END
INSERT INTO Student(StudentName) VALUES(@StudentName)
IF @@Error = 0
BEGIN
SET @StudentID = @@Identity
RETURN
END
RAISERROR(‘Insert operation failed. Error Code - %d’, @@Error )
GO
The above stored procedure will insert a new student name into the student table
and return the studentid as the output parameter.

GLOBAL Variables
I found the following variables as most useful.
@@Error - Returns the error number for the last Transact-SQL statement executed.
@@FETCH_STATUS - Returns the status of the last cursor FETCH statement issued
against any cursor currently opened by the connection.
@@IDENTITY - Returns the last-inserted identity value.
@@ROWCOUNT - Returns the number of rows affected by the last statement.

Indexes
Indexes are created for the fast execution of Query. Unnecessary creation of indexes
will slow down your application, so don’t create indexes until you finished creating
the queries. INSERT statement will work faster if there is no index at all. To learn
you have to read about SCAN and SEEK operations first. SCAN – It is an operation
which will move sequentially from one record to another record to compare a value.
SEEK – This operation will move to the record position based on the group of record.
SEEK operations are faster for query execution.
Only one index per query can be used by SQL server. So be careful about creating
index for a query.
Execution Plan
You can view execution plan using SQL query analyzer. Using the execution plan you
can redesign the table or create, modify drop indexes.
Table Hints
Well you learned to create indexes and created an index that may made your query
execution faster because SQL server will pick up indexes automatically for query
execution. But in few cases, if there are more than one index on table you have
specify the index you want to use in the query, otherwise SQL server may pickup
some other index which may not give the full performance. (It don’t mean SQL
server will randomly pickup the indexes but meant SQL server may miss to pickup
the index which you want).
The following query will use the index IX_EmployeeEmpName defined on the
Employee table extract the employee names.
SELECT EmpName
FROM Employee WITH (INDEX (IX_EmployeeEmpName))
WHERE EmpName LIKE ‘Test%’
The following query will use the index IX_EmployeeEmpId to extract employee name.
SELECT EmpName
FROM Employee WITH (NOLOCK, INDEX (IX_EmployeeEmpId))
WHERE EmpId = 10000

DTS (Data Transformation Services)
Data Transformation Services are used to transfer and transform data from one
datasource to another datasoure in the same server or another server.
Some applications of DTS
1) Copying data from one database to another database
2) Copying data structure from one database to another database
3) Migrating data from other datasources such as Flat File, Excel, Access, Oracle
to SQL Server.
4) Migrating data from SQL server to other datasources.
Since the DTS topic is huge you cannot get to know what it is until you do actual task
that is related to DTS. So try the following things and you will get to know something
about DTS. Create an Excel file with the following columns EmployeeName, Address,
DateOfJoin, PhoneNumber, MaritialStatus, and Department. Fill this excel sheet with
some meaningful information and then try to transfer the contents from Excel to SQL
using DTS Import/Export Wizard.
You can also create DTS packages which can be scheduled to run at future. DTS
package programming allows mapping source fields to different destination fields and
also provides error control.

SQL Profiler
SQL Profiler is a handy tool used to analyze what is happening inside and outside of
a SQL Server. Simply it is a tool which extracts the log information from SQL server.
These logs will help you debug applications, optimize queries, redesign database
structure.
SQL Profiler is a powerful which has no alternate tool to extract trace information
from the SQL Server. You can change the trace settings of SQL Server by using SQL
Profiler or by using system stored procedures such as sp_trace_create,
sp_trace_setfilter, sp_trace_generateevent, sp_trace_setstatus and
sp_trace_setevent. SQL server writes one record in the tempdb.TraceQueue table for
each trace it makes.
Running SQL Profiler is very easy, Click SQL profiler from the SQL Server group
menu. Then select New Trace from the file menu. It will open the connection dialog
box. Type your server name, user name and password. Now you have to set options
for your profile. The options include what events you want to trace, what columns
you want in the output, filters if any.
For example you may want to see all applications that access a table. Then you
have add the Object:Opened event and also add ApplicationName in the Data
Columns. This trace will track all the events raised against all objects, but we want
only a particular table in a specific database, to do that you have to add a filter to
this trace. So add two filters ObjectName Like and DataBaseName Like. Then
click Run. It will start tracing after that.

DBCC
Database Console Commands are referred as DBCC. DBCC contains some special
commands through which you can accomplish certain DB operations which is not
possible through normal SQL commands.
For example you can check the allocation and structural integrity of all the objects in
the specified database using the DBCC CHECKDB command.
I have used the following commands and found they are very useful.
DBCC INDEXDEFRAG - Defragments clustered and secondary indexes of the specified
table or view.
DBCC SHRINKDATABASE - Shrinks the size of the data files in the specified
database. This command is very useful to reduce the size of the data files and log
files.
DBCC INPUTBUFFER - Displays the last statement sent from a client. You have to
pass the spid of the client to this command. To get the client id use the following
query – SELECT * FROM sysprocesses
DBCC CHECKIDENT - Checks the current identity value for the specified table and, if
needed, corrects the identity value.
The End

About Me
I am not a DBA. But I have some working knowledge in SQL. Although my interest is
in system side, I always want to make some documentations/articles. And I hope I
will move into system side programming soon so I make this article as my
contribution to application side programming. I like to give my thanks to
Mathiazhagan and Chandra Babu for giving early feedback on this article.
You can get the latest version of this document from the following URL:

Introduction to sql server

More Related Content

What's hot (18)

Viewers also liked (8)

Similar to Introduction to sql server (20)

Introduction to sql server