1 data types

Working with MSSQL Server
Data Types
Ram Kedem

Copyright 2014 © Ram Kedem. All rights reserved. Not to be reproduced without written consent
Data Types Introduction
• Data types determine what can be stored
• Constrains the type of data that an object can hold
• Provides limits on the range of values
• Critical to choose appropriate data type
• It constrains the type of data an object can hold. This means that
you can't put text in a numeric object, etc.
• It rovides limits on the range of values. If you know it's an integer,
you can't try to put 2304892309232323423 in it.
• It Assists with query optimization. If SQL Server knows a column is
an integer, it can build a different query plan to one where the
value might also be a name.
• It provides a level of self-documentation. If all columns were
variant, you would need to separately document the type of data
stored in them to know how to work with them.

Exact Numeric Data Types - Integer
DROP TABLE int_tab ;
CREATE TABLE int_tab
(c_tinyint tinyint, -- (0 to 255),
-- Cannot store negative numbers
c_smallint smallint, -- (-32768 to 32767)
c_int int, -- (-2147483648 to 2147483647)
c_bigint bigint -- (-2^63 to 2^63 – 1)
)
INSERT INTO int_tab
VALUES (1.5 , 1.6 , 1.7 , 1.8)
SELECT * FROM int_tab ;
DELETE FROM int_tab ;

Exact Numeric Data Types -Decimal
• decimal and numeric are functionally equivalent.
• we normally write decimal(p,s)
• p (precision) is a value from 1 to 38 with a default of 18 and
• S (scale) is the number of decimal places and can be any value
from 0 up to the P value.
• The default scale is zero.

Exact Numeric Data Types - Decimal
DROP TABLE dec_tab
CREATE TABLE dec_tab
(c_decimal_1 decimal, -- Default is decimal(18)
c_decimal_2 decimal (8,2),
-- c_decimal_3 decimal (8,9),
-- Error : The scale (9) for column 'c_decimal_3' must
-- be within the range 0 to 8.
c_numeric_1 numeric,
c_numeric_2 numeric (8,2))
INSERT INTO dec_tab
VALUES (1.5 , 1.6 , 1.7 , 1.8)
INSERT INTO dec_tab
VALUES (1, 1 ,123456789123456789 ,1)
SELECT * FROM dec_tab ;

Exact Numeric Data Types -Currency
• Money and Smallmoney have a fixed scale of 4
• Typically decimal would be a better (and more standard)
choice
DROP TABLE curr_tab
CREATE TABLE curr_tab
(c_smallmoney smallmoney, -- Maximum precision is 6, fixed scale of 4.
c_money money, -- Maximum precision is 15, fixed scale of 4.
c_bit bit) -- values of 1, 0, or NULL
INSERT INTO curr_tab
VALUES (1.12345 , 1.12345 , 1)
SELECT * FROM curr_tab ;

Approximate Numeric Data Types
• You should try to avoid the FLOAT datatype whenever
possible, and opt for the more versatile, and precise, DECIMAL
or NUMERIC datatypes instead
• float is from float(1) to float(53) and defaults to float(53)
• The ISO synonym for real is float(24).

Approximate Numeric Data Types
DROP TABLE approx_tab
CREATE TABLE approx_tab
(c_real_1 real,
c_real_2 real,
-- c_real_3 real(4), -- Error - Cannot specify a column width on data type real.
c_float_1 float(3),
c_float_2 float)
INSERT INTO approx_tab
VALUES (54 , 3.1 , 1.3, 1.4)
SELECT * FROM approx_tab
SELECT (0 + c_real_2 + c_real_1) - c_real_1 - c_real_2
FROM approx_tab
DECLARE @Float1 float, @Float2 float, @Float3 float, @Float4 float;
SET @Float1 = 54;
SET @Float2 = 0.03;
SET @Float3 = 0 + @Float1 + @Float2;
SELECT @Float3 - @Float1 - @Float2 AS "Should be 0";

Exact VS Approximate
• A grand choice for “exact” numeric data, is when dealing with
numbers that have a fixed number of decimal places and
represent an exact amount, such as monetary units.
• But if you are dealing with scientific data, that is usually
derived from some measurement and hence by definition an
approximation of reality (since there’s no way to measure with
unlimited precision) , floating point data is an excellent choice

Dates
CREATE TABLE dates_tab
(c_date date,
-- 0001-01-01 to 9999-12-31
-- Accuracy of 1 day
c_datetime2 datetime2,
-- 0001-01-01 to 9999-12-31
-- Accuracy of 100 nanosecond
-- DateTime2 is the new Data Type introduced in Sql Server 2008
-- Microsoft Suggests to use this new Data Type for new work instead of DateTime
c_datetime datetime,
-- 1753-01-01 to 9999-12-31
-- Accuracy of 3.33 milliseconds
c_datetimeoffset datetimeoffset ,
-- 0001-01-01 to 9999-12-31
-- Plus timezone
c_smalldatetime smalldatetime,
-- 1900-01-01 to 2079-06-06
-- Accuracy of 1 minute
c_time time
-- 00:00:00.0000000 to 23:59:59.9999999
)

Dates
INSERT INTO dates_tab
VALUES (GETDATE() , GETDATE() , GETDATE() ,
GETDATE() , GETDATE() , GETDATE() )
SELECT * FROM dates_tab
SELECT c_date AS 'Date' ,
DATALENGTH ( c_date ) AS 'Date Size',
c_datetime2 AS 'Datetime2' ,
DATALENGTH ( c_datetime2 ) AS 'Datetime2 Size',
c_datetime AS 'datetime',
DATALENGTH ( c_datetime ) AS 'datetime Size',
c_datetimeoffset AS 'datetimeoffset',
DATALENGTH ( c_datetimeoffset )AS 'datetimeoffset Size',
c_smalldatetime AS 'smalldatetime',
DATALENGTH ( c_smalldatetime ) AS 'smalldatetime Size',
c_time AS 'time',
DATALENGTH ( c_time ) AS 'time Size'
FROM dates_tab

Working with Character Data
Unicode / Non Unicode
• Unicode
• Is a worldwide character-encoding standard
• Simplifies software localization
• Improves multilingual character processing
• Is implemented in SQL Server as double-byte for Unicode types
Requires N prefix on constants
• Uses LEN() to return number of characters, DATALENGTH() to
return the number of bytes

DROP TABLE dbo.testcharacter ;
CREATE TABLE dbo.TestCharacter
(
id int NOT NULL IDENTITY(1,1),
NonUnicodeData varchar(10),
UnicodeData nvarchar(10)
);
INSERT dbo.TestCharacter (NonUnicodeData,UnicodeData)
VALUES ('Hello',N'Hello'),
('‫,'שלום‬N'‫)'שלום‬

SELECT * FROM dbo.TestCharacter;
SELECT nonUnicodeData ,
LEN(nonUnicodeData) AS 'number of characters' ,
DATALENGTH(nonUnicodeData) AS 'number of bytes'
FROM dbo.TestCharacter
WHERE nonUnicodeData = 'hello' ;
SELECT UnicodeData ,
LEN(UnicodeData) AS 'number of characters' ,
DATALENGTH(UnicodeData) AS 'number of bytes'
WHERE UnicodeData = N'‫'שלום‬ ;
SELECT UnicodeData ,
LEN(UnicodeData) AS 'number of characters' ,
DATALENGTH(UnicodeData) AS 'number of bytes'
WHERE UnicodeData = 'hello' ;

Character Data Types
DROP TABLE char_tab ;
CREATE TABLE char_tab
(c_char char(5), -- maximum 8000
c_nchar nchar(20), -- maximum 4000
c_varchar varchar, -- maximum 8000
c_nvarchar nvarchar, -- maximum 4000
c_varchar_max varchar(max), -- maximum 2G
c_nvarchar_max nvarchar(max), -- maximum 2G
c_text text, -- deprecated, use varchar(max)
c_ntext ntext) -- deprecated, use nvarchar(max)

Understanding Collations
• Collation refers to a set of rules that determine how data is
sorted and compared.
• Character data is sorted using rules that define the correct
character sequence, with options for specifying case-
sensitivity, accent marks, kana character types and character
width.

-- The CS and AS values mean "case-sensitive" and "accent-
sensitive".
DROP TABLE collate_tab
GO
CREATE TABLE Collate_tab
(person_id int identity (1,1),
person_name varchar(25));
GO
INSERT INTO Collate_tab
VALUES ('Ram') , ('ram') , ('RAM') , ('rAM');
SELECT * FROM Collate_tab ;

-- By defualt not case sensitive
SELECT * FROM Collate_tab WHERE person_name = 'ram';
-- Change to case sensitive in the SQL Statement level
SELECT *
FROM Collate_tab
WHERE person_name COLLATE Latin1_General_CS_AS= 'ram';
-- Change to case sensitive in the Table column level
ALTER TABLE Collate_tab
ALTER COLUMN person_name VARCHAR(25)
COLLATE Latin1_General_CS_AS
SELECT *
FROM Collate_tab
WHERE person_name = 'ram';
-- Change in the DB level
-- ALTER DATABASE adventureworks collate SQL_Latin1_General_CP1_CI_AS

CAST and CONVERT
-- Using cast
--------------
-- CAST is supported by SQL standards
SELECT 'The date is : ' + CAST(GETDATE() AS VARCHAR(20))
-- Using Covert
----------------
-- CONVERT is a more powerful SQL Server extension to the SQL language
SELECT 'The date is : ' + CONVERT(VARCHAR(20) ,GETDATE() , 104)
USE tempdb;
GO
-- Step 2: Execute statements to see different formats
-- of a date conversion
SELECT CAST(SYSDATETIME() as varchar(20)) as CastDate,
CONVERT(varchar(20), SYSDATETIME(), 110) as USA,
CONVERT(varchar(20), SYSDATETIME(), 104) as German,
CONVERT(varchar(20), SYSDATETIME(), 105) as Italian,
CONVERT(varchar(20), SYSDATETIME(), 112) as ISO,
CONVERT(varchar(20), SYSDATETIME(), 114) as [Time];
GO

Parse
-- http://go.microsoft.com/fwlink/?LinkID=233801
SELECT PARSE('2010 13 December' AS datetime2 USING 'en-US');
SELECT PARSE('$345.98' AS money USING 'en-US') AS Result;
-- error
SELECT PARSE ('NotLikely' AS decimal USING 'en-US') AS result
-- null
SELECT TRY_PARSE ('NotLikely' AS decimal USING 'en-US') AS result ;
SELECT
CASE WHEN TRY_PARSE('NotLikely' AS decimal
USING
'sr-Latn-CS') IS NULL
THEN 'Not Decimal'
ELSE 'Decimal'
END AS Result ;

Implicit Conversion
DECLARE @int1 int = 1;
DECLARE @int2 int = 2;
DECLARE @char char(1) = 3;
SELECT @int1 + @int2 ;
SELECT @int2 * @char
GO

ROWVERSION
• Rowversion is the synonym for timestamp in SQL Server 2005
and SQL Server 2008.
• Timestamp has been deprecated and replaced by rowversion.
• Rowversion should be used for all new work. timestamp does
not contain any value related to time.

ROWVERSION
USE tempdb;
GO
-- Step 2: Create and populate a table that uses the rowversion data type
CREATE TABLE dbo.Orders
(
OrderID int NOT NULL
IDENTITY(1,1),
OrderDate date NOT NULL,
SalespersonID int NULL,
Concurrency rowversion NOT NULL
);
GO
-- Step 3: Add 2 new rows to dbo.Opportunity
INSERT dbo.Orders (OrderDate)
VALUES (SYSDATETIME()),
(SYSDATETIME());
GO

ROWVERSION
-- Step 4: Show that the rowversion column was populated automatically
SELECT * FROM dbo.Orders;
GO
-- Step 5: Try to update the rowversion column. This will fail as
-- you cannot update a rowversion column directly.
UPDATE dbo.Orders
SET Concurrency = Concurrency +1
WHERE OrderID =1;
GO
-- Step 6: Try to insert an explicit value for the rowversion column.
-- This will fail as you cannot insert into a rowversion
-- column directly.
INSERT dbo.Orders (OrderDate, Concurrency)
VALUES (SYSDATETIME(), 0x00000000000007E4);
GO

ROWVERSION
-- Step 7: Update another column for OrderID 1
SELECT * FROM dbo.Orders
WHERE OrderID = 1;
UPDATE dbo.Orders
SET SalespersonID = 36
WHERE OrderID =1;
SELECT * FROM dbo.Orders
WHERE OrderID = 1;
GO

Alias Data Types
• The main use of these alias types is in maintaining consistency
across development.
CREATE TYPE ProductNumber
FROM nvarchar(20) NOT NULL;
GO
CREATE TABLE Production.ProductConversion
( ProductConversionID int IDENTITY(1,1),
FromProduct ProductNumber,
ToProduct ProductNumber
);

Computed Columns
• Computed columns are actually not physically created on
the related database table unless they are defined as
"Persisted" computed columns.
• You can think of computed columns as virtual columns.
They are not physically stored in the related sql database
table.
• A non-persisted computed column is calculated every
time a SELECT operation occurs on the column.
• A persisted computed column is calculated when the
data in the row is inserted or updated

Computed Columns
-- Step 1: Open a new query window to the tempdb database
USE tempdb;
GO
-- Step 2: Create a table with a persisted computed column
DROP TABLE pet ;
GO
CREATE TABLE Pet
(PetID int IDENTITY(1,1) PRIMARY KEY,
PetName nvarchar(30) NOT NULL,
PetType nvarchar(10) NOT NULL,
DateOfBirth date NOT NULL,
YearOfBirth AS DATEPART(year,DateOfBirth) PERSISTED
);
GO
-- Step 3: Insert some data into the table
INSERT Pet (PetName, PetType, DateOfBirth)
VALUES ('Brutus','DOG','2007-12-31'),
('Bruno','DOG','2006-01-01'),
('Tweetie','BIRD','2005-04-02');
GO
-- Step 4: Query the table
SELECT * FROM Pet;
GO
GO

Computed Columns
-- Demo 2 - Using CASE ()
-------------------------
CREATE TABLE emps
(id int,
sal int,
grade_level as (CASE WHEN sal BETWEEN 0 AND 5000 THEN 'a' ELSE 'b' END ))
INSERT INTO emps VALUES (1, 1)
SELECT * FROM emps
-- Demo 3 - Combination of different columns
--------------------------------------------
DROP TABLE emps
CREATE TABLE emps
(id int,
sal_per_hour int,
total_hours int,
total_sal AS sal_per_hour * total_hours PERSISTED)
INSERT INTO emps VALUES (1, 25, 180)
SELECT * FROM emps
UPDATE emps SET sal_per_hour = 40 ;

Computed Columns
-- Demo 4 - Computed column in another computed column
------------------------------------------------------
DROP TABLE emps
CREATE TABLE emps
(id int,
birthdate datetime,
age as datediff(yyyy, birthdate , getdate()),
emp_rank as (CASE WHEN age BETWEEN 0 AND 30 THEN 'Begginer' ELSE 'Senior' END) )

1 data types

More Related Content

What's hot (20)

Viewers also liked (10)

Similar to 1 data types (20)

More from Ram Kedem (20)

Recently uploaded (20)

1 data types