Getting started with Microsoft SQL Server
INSERT / SELECT / UPDATE / DELETE: the basics of Data Manipulation Language
Data Manipulation Language (DML for short) includes operations such as
-- Create a table HelloWorld CREATE TABLE HelloWorld ( Id INT IDENTITY, Description VARCHAR(1000) ) -- DML Operation INSERT, inserting a row into the table INSERT INTO HelloWorld (Description) VALUES ('Hello World') -- DML Operation SELECT, displaying the table SELECT * FROM HelloWorld -- Select a specific column from table SELECT Description FROM HelloWorld -- Display number of records in the table SELECT Count(*) FROM HelloWorld -- DML Operation UPDATE, updating a specific row in the table UPDATE HelloWorld SET Description = 'Hello, World!' WHERE Id = 1 -- Selecting rows from the table (see how the Description has changed after the update?) SELECT * FROM HelloWorld -- DML Operation - DELETE, deleting a row from the table DELETE FROM HelloWorld WHERE Id = 1 -- Selecting the table. See table content after DELETE operation SELECT * FROM HelloWorld
In this script we’re creating a table to demonstrate some basic queries.
The following examples are showing how to query tables:
USE Northwind; GO SELECT TOP 10 * FROM Customers ORDER BY CompanyName
will select the first 10 records of the
Customer table, ordered by the column
CompanyName from the database
Northwind (which is one of Microsoft’s sample databases, it can be downloaded from here):
Use Northwind; changes the default database for all subsequent queries.
You can still reference the database by using the fully qualified syntax in the form of [Database].[Schema].[Table]:
SELECT TOP 10 * FROM Northwind.dbo.Customers ORDER BY CompanyName SELECT TOP 10 * FROM Pubs.dbo.Authors ORDER BY City
This is useful if you’re querying data from different databases. Note that
dbo, specified “in between” is called a schema and needs to be specified while using the fully qualified syntax. You can think of it as a folder within your database.
dbo is the default schema. The default schema may be omitted. All other user defined schemas need to be specified.
If the database table contains columns which are named like reserved words, e.g.
Date, you need to enclose the column name in brackets, like this:
-- descending order SELECT TOP 10 [Date] FROM dbo.MyLogTable ORDER BY [Date] DESC
The same applies if the column name contains spaces in its name (which is not recommended). An alternative syntax is to use double quotes instead of square brackets, e.g.:
-- descending order SELECT top 10 "Date" from dbo.MyLogTable order by "Date" desc
is equivalent but not so commonly used. Notice the difference between double quotes and single quotes: Single quotes are used for strings, i.e.
-- descending order SELECT top 10 "Date" from dbo.MyLogTable where UserId='johndoe' order by "Date" desc
is a valid syntax. Notice that T-SQL has a
N prefix for NChar and NVarchar data types, e.g.
SELECT TOP 10 * FROM Northwind.dbo.Customers WHERE CompanyName LIKE N'AL%' ORDER BY CompanyName
returns all companies having a company name starting with
% is a wild card, use it as you would use the asterisk in a DOS command line, e.g.
DIR AL*). For
LIKE, there are a couple of wildcards available, look here to find out more details.
Joins are useful if you want to query fields which don’t exist in one single table, but in multiple tables. For example: You want to query all columns from the
Region table in the
Northwind database. But you notice that you require also the
RegionDescription, which is stored in a different table,
Region. However, there is a common key,
RgionID which you can use to combine this information in a single query as follows (
Top 5 just returns the first 5 rows, omit it to get all rows):
SELECT TOP 5 Territories.*, Regions.RegionDescription FROM Territories INNER JOIN Region ON Territories.RegionID=Region.RegionID ORDER BY TerritoryDescription
will show all columns from
Territories plus the
RegionDescription column from
Region. The result is ordered by
When your query requires a reference to two or more tables, you may find it useful to use a Table Alias. Table aliases are shorthand references to tables that can be used in place of a full table name, and can reduce typing and editing. The syntax for using an alias is:
<TableName> [as] <alias>
as is an optional keyword. For example, the previous query can be rewritten as:
SELECT TOP 5 t.*, r.RegionDescription FROM Territories t INNER JOIN Region r ON t.RegionID = r.RegionID ORDER BY TerritoryDescription
Aliases must be unique for all tables in a query, even if you use the same table twice. For example, if your Employee table included a SupervisorId field, you can use this query to return an employee and his supervisor’s name:
SELECT e.*, s.Name as SupervisorName -- Rename the field for output FROM Employee e INNER JOIN Employee s ON e.SupervisorId = s.EmployeeId WHERE e.EmployeeId = 111
As we have seen before, a Join adds columns from different table sources. But what if you want to combine rows from different sources? In this case you can use a UNION. Suppose you’re planning a party and want to invite not only employees but also the customers. Then you could run this query to do it:
SELECT FirstName+' '+LastName as ContactName, Address, City FROM Employees UNION SELECT ContactName, Address, City FROM Customers
It will return names, addresses and cities from the employees and customers in one single table. Note that duplicate rows (if there should be any) are automatically eliminated (if you don’t want this, use a
UNION ALL instead). The column number, column names, order and data type must match across all the select statements that are part of the union - this is why the first SELECT combines
LastName from Employee into
It can be useful, if you need to deal with temporary data (especially in a stored procedure), to use table variables: The difference between a “real” table and a table variable is that it just exists in memory for temporary processing.
DECLARE @Region TABLE ( RegionID int, RegionDescription NChar(50) )
creates a table in memory. In this case the
@ prefix is mandatory because it is a variable. You can perform all DML operations mentioned above to insert, delete and select rows, e.g.
INSERT INTO @Region values(3,'Northern') INSERT INTO @Region values(4,'Southern')
But normally, you would populate it based on a real table like
INSERT INTO @Region SELECT * FROM dbo.Region WHERE RegionID>2;
which would read the filtered values from the real table
dbo.Region and insert it into the memory table
@Region - where it can be used for further processing. For example, you could use it in a join like
SELECT * FROM Territories t JOIN @Region r on t.RegionID=r.RegionID
which would in this case return all
More detailed information can be found here. Temporary tables are discussed here, if you are interested to read more about that topic.
NOTE: Microsoft only recommends the use of table variables if the number of rows of data in the table variable are less than 100. If you will be working with larger amounts of data, use a temporary table, or temp table, instead.
SELECT all rows and columns from a table
SELECT * FROM table_name
Using the asterisk operator
* serves as a shortcut for selecting all the columns in the table. All rows will also be selected because this
SELECT statement does not have a
WHERE clause, to specify any filtering criteria.
This would also work the same way if you added an alias to the table, for instance
e in this case:
SELECT * FROM Employees AS e
Or if you wanted to select all from a specific table you can use the alias + " .* “:
SELECT e.*, d.DepartmentName FROM Employees AS e INNER JOIN Department AS d ON e.DepartmentID = d.DepartmentID
Database objects may also be accessed using fully qualified names:
SELECT * FROM [server_name].[database_name].[schema_name].[table_name]
This is not necessarily recommended, as changing the server and/or database names would cause the queries using fully-qualified names to no longer execute due to invalid object names.
Note that the fields before
table_name can be omitted in many cases if the queries are executed on a single server, database and schema, respectively. However, it is common for a database to have multiple schema, and in these cases the schema name should not be omitted when possible.
SELECT * in production code or stored procedures can lead to problems later on (as new columns are added to the table, or if columns are rearranged in the table), especially if your code makes simple assumptions about the order of columns, or number of columns returned. So it’s safer to always explicitly specify column names in SELECT statements for production code.
SELECT col1, col2, col3 FROM table_name
UPDATE Specific Row
UPDATE HelloWorlds SET HelloWorld = 'HELLO WORLD!!!' WHERE Id = 5
The above code updates the value of the field “HelloWorld” with “HELLO WORLD!!!” for the record where “Id = 5” in HelloWorlds table.
Note: In an update statement, It is advised to use a “where” clause to avoid updating the whole table unless and until your requirement is different.
Comments in code
Transact-SQL supports two forms of comment writing. Comments are ignored by the database engine, and are meant for people to read.
Comments are preceded by
-- and are ignored until a new line is encountered:
-- This is a comment SELECT * FROM MyTable -- This is another comment WHERE Id = 1;
Slash star comments begin with
/* and end with
*/. All text between those delimiters is considered as a comment block.
/* This is a multi-line comment block. */ SELECT Id = 1, [Message] = 'First row' UNION ALL SELECT 2, 'Second row' /* This is a one liner */ SELECT 'More';
Slash star comments have the advantage of keeping the comment usable if the SQL Statement loses new line characters. This can happen when SQL is captured during troubleshooting.
Slash star comments can be nested and a starting
/* inside a slash star comment needs to be ended with a
*/ to be valid. The following code will result in an error
/* SELECT * FROM CommentTable WHERE Comment = '/*' */
The slash star even though inside the quote is considered as the start of a comment. Hence it needs to be ended with another closing star slash. The correct way would be
/* SELECT * FROM CommentTable WHERE Comment = '/*' */ */
DELETE All Rows
DELETE FROM Helloworlds
This will delete all the data from the table. The table will contain no rows after you run this code. Unlike
DROP TABLE, this preserves the table itself and its structure and you can continue to insert new rows into that table.
Another way to delete all rows in table is truncate it, as follow:
TRUNCATE TABLE HelloWords
Difference with DELETE operation are several:
- Truncate operation doesn’t store in transaction log file
- If exists
IDENTITYfield, this will be reset
- TRUNCATE can be applied on whole table and no on part of it (instead with
DELETEcommand you can associate a
Restrictions Of TRUNCATE
- Cannot TRUNCATE a table if there is a
- If the table is participated in an
- If the table is published by using
- It will not fire any TRIGGER defined in the table
Display a message to the output console. Using SQL Server Management Studio, this will be displayed in the messages tab, rather than the results tab:
PRINT 'Hello World!';
Select rows that match a condition
Generally, the syntax is:
SELECT <column names> FROM <table name> WHERE <condition>
SELECT FirstName, Age FROM Users WHERE LastName = 'Smith'
Conditions can be complex:
SELECT FirstName, Age FROM Users WHERE LastName = 'Smith' AND (City = 'New York' OR City = 'Los Angeles')
UPDATE All Rows
A simple form of updating is incrementing all the values in a given field of the table. In order to do so, we need to define the field and the increment value
The following is an example that increments the
Score field by 1 (in all rows):
UPDATE Scores SET score = score + 1
This can be dangerous since you can corrupt your data if you accidentally make an UPDATE for a specific Row with an UPDATE for All rows in the table.
Retrieve Basic Server Information
Returns the version of MS SQL Server running on the instance.
Returns the name of the MS SQL Server instance.
Returns the name of the Windows service MS SQL Server is running as.
Returns the physical name of the machine where SQL Server is running. Useful to identify the node in a failover cluster.
SELECT * FROM fn_virtualservernodes();
In a failover cluster returns every node where SQL Server can run on. It returns nothing if not a cluster.
TRUNCATE TABLE Helloworlds
This code will delete all the data from the table Helloworlds. Truncate table is almost similar to
Delete from Table code. The difference is that you can not use where clauses with Truncate. Truncate table is considered better than delete because it uses less transaction log spaces.
Note that if an identity column exists, it is reset to the initial seed value (for example, auto-incremented ID will restart from 1). This can lead to inconsistency if the identity columns is used as a foreign key in another table.
Create new table and insert records from old table
SELECT * INTO NewTable FROM OldTable
Creates a new table with structure of old table and inserts all rows into the new table.
- You cannot specify a table variable or table-valued parameter as the new table.
- You cannot use SELECT…INTO to create a partitioned table, even when the source table is partitioned. SELECT…INTO does not use the partition scheme of the source table; instead, the new table is created in the default filegroup. To insert rows into a partitioned table, you must first create the partitioned table and then use the INSERT INTO…SELECT FROM statement.
- Indexes, constraints, and triggers defined in the source table are not transferred to the new table, nor can they be specified in the SELECT…INTO statement. If these objects are required, you can create them after executing the SELECT…INTO statement.
- Specifying an ORDER BY clause does not guarantee the rows are inserted in the specified order. When a sparse column is included in the select list, the sparse column property does not transfer to the column in the new table. If this property is required in the new table, alter the column definition after executing the SELECT…INTO statement to include this property.
- When a computed column is included in the select list, the corresponding column in the new table is not a computed column. The values in the new column are the values that were computed at the time SELECT…INTO was executed.
Using Transactions to change data safely
Whenever you change data, in a Data Manipulation Language(DML) command, you can wrap your changes in a transaction. DML includes
DELETE. One of the ways that you can make sure that you’re changing the right data would be to use a transaction.
DML changes will take a lock on the rows affected. When you begin a transaction, you must end the transaction or all objects being changed in the DML will remain locked by whoever began the transaction. You can end your transaction with either
ROLLBACK returns everything within the transaction to its original state.
COMMIT places the data into a final state where you cannot undo your changes without another DML statement.
--Create a test table USE [your database] GO CREATE TABLE test_transaction (column_1 varchar(10)) GO INSERT INTO dbo.test_transaction ( column_1 ) VALUES ( 'a' ) BEGIN TRANSACTION --This is the beginning of your transaction UPDATE dbo.test_transaction SET column_1 = 'B' OUTPUT INSERTED.* WHERE column_1 = 'A' ROLLBACK TRANSACTION --Rollback will undo your changes --Alternatively, use COMMIT to save your results SELECT * FROM dbo.test_transaction --View the table after your changes have been run DROP TABLE dbo.test_transaction
- This is a simplified example which does not include error handling. But any database operation can fail and hence throw an exception. Here is an example how such a required error handling might look like. You should never use transactions without an error handler, otherwise you might leave the transaction in an unknown state.
- Depending on the isolation level, transactions are putting locks on the data being queried or changed. You need to ensure that transactions are not running for a long time, because they will lock records in a database and can lead to deadlocks with other parallel running transactions. Keep the operations encapsulated in transactions as short as possible and minimize the impact with the amount of data you’re locking.
Getting Table Row Count
The following example can be used to find the total row count for a specific table in a database if
table_name is replaced by the the table you wish to query:
SELECT COUNT(*) AS [TotalRowCount] FROM table_name;
It is also possible to get the row count for all tables by joining back to the table’s partition based off the tables’ HEAP (index_id = 0) or cluster clustered index (index_id = 1) using the following script:
SELECT [Tables].name AS [TableName], SUM( [Partitions].[rows] ) AS [TotalRowCount] FROM sys.tables AS [Tables] JOIN sys.partitions AS [Partitions] ON [Tables].[object_id] = [Partitions].[object_id] AND [Partitions].index_id IN ( 0, 1 ) --WHERE [Tables].name = N'table name' /* uncomment to look for a specific table */ GROUP BY [Tables].name;
This is possible as every table is essentially a single partition table, unless extra partitions are added to it. This script also has the benefit of not interfering with read/write operations to the tables rows'.