1. Working with SQL Server 2016 Data Types
20761B
6: Working with SQL Server 2016 Data Types
Module 6
All demonstrations in this module use a Microsoft Azure® SQL Database running a copy of the AdventureWorksLT database.
Before attempting to run these demos, ensure you have a copy of AdventureWorksLT running on an Azure instance.
For detailed steps on creating a copy of the AdventureWorksLT database in Azure see: D:\Creating an AdventureWorks Database on Azure.docx.
Working with SQL Server 2016 Data Types

2. Lesson 1: Introducing SQL Server 2016 Data Types
20761B
Lesson 1: Introducing SQL Server 2016 Data Types
6: Working with SQL Server 2016 Data Types
Demonstration: SQL Server Data Types
Note that character and date/time datatypes will be covered in later lessons.

3. SQL Server supplies built-in data types
SQL Server Data Types
6: Working with SQL Server 2016 Data Types
SQL Server associates columns, expressions, variables and parameters with data types
Data types determine the kind of data that can be held in a column or variable
Integers, characters, dates, decimals, binary strings, and so on
SQL Server supplies built-in data types
Developers can also define custom data types
Point out that approximate numeric data types (also referred to as floating point numbers) are typically used in scientific applications; business applications most commonly use exact numeric types.
SQL Server Data Type Categories
Exact numeric
Unicode character strings
Approximate numeric
Binary strings
Date and time
Other
Character strings

4. Exact Numeric Data Types
20761B
Numeric Data Types
6: Working with SQL Server 2016 Data Types
Exact Numeric Data Types
Data Type
Range
Storage (bytes)
tinyint
0 to 255 1
smallint
-32,768 to 32,768 2
int
231 (-2,147,483,648) to (2,147,483,647) 4
bigint
(+/- 9 quintillion) 8
bit
1, 0 or NULL
decimal/numeric
through 1038 – 1 when maximum precision is used
5-17
money
-922,337,203,685, to 922,337,203,685,
smallmoney
-214, to 214,
Note that not all content from the tables is printed in the workbook—use the links provided to show the references in the SQL Server 2016 Technical Documentation.
Point out that numeric data types should be selected on the data they will need to store; storage considerations should be secondary. The int data type has often been the default choice amongst the integer data types, but might not always be most suitable.
Decimal precision: The maximum total number of decimal digits that can be stored, both to the left and to the right of the decimal point. The precision must be a value from 1 through the maximum precision of 38, with the default set at 18.
Decimal scale: The maximum number of decimal digits that can be stored to the right of the decimal point.
Float mantissa value: The default value for the optional float mantissa (the number of bits used to store the mantissa of the float number in scientific notation) is 53.
float(53) is a synonym for double precision (the double data type in various programming languages including C# and VB.Net).
For more information, see the following topics in the SQL Server 2016 Technical Documentation:
Decimal and Numeric (Transact SQL)
Precision, Scale, and Length (Transact-SQL)
Data Types (Transact-SQL)
Float and Real (Transact-SQL)

5. Binary String Data Types
6: Working with SQL Server 2016 Data Types
Binary string data types
The image data type is also a binary string type but is marked for removal in a future version of SQL Server; varbinary(max) should be used instead
Data Type
Range
Storage (bytes)
binary(n)
1 to 8000 bytes
n bytes
varbinary(n)
n bytes + 2
varbinary(max)
1 to 2.1 billion (approx.) bytes
See the binary and varbinary (Transact-SQL) topic in the SQL Server 2016 Technical Documentation.
Note that all data, regardless of the data type, is held in a binary format in SQL Server. Unlike other data types, the binary string data types store and retrieve binary data without modification or implicit formatting.

6. Other Data Types Data Type Range Storage (bytes) Remarks xml 0-2 GB
6: Working with SQL Server 2016 Data Types
Data Type
Range
Storage (bytes)
Remarks
xml
0-2 GB
Stores XML in native hierarchical structure
uniqueidentifier
Auto-generated 16
Globally unique identifier (GUID)
hierarchyid
n/a
Depends on content
Represents position in a hierarchy
rowversion 8
Previously called timestamp
geometry
Shape definitions in Euclidian geometry
geography
Shape definitions in round-earth geometry
sql_variant
bytes
Can store data of various other data types in the same column
cursor
Not a storage datatype—used for cursor operations
table
Not a storage data type—used for query operations
Rowversion: note that, in previous versions of SQL Server, this data type was called timestamp, even though no date/time information was stored.
For more information on XML, see Course
Some of these types (cursors, sql_variant, and so on) are only listed for completeness. Don't get bogged down in details here.
There's no need to get into design discussions at this stage, but you might wish to point out that the use of sql_variant probably means poor analysis of the problem and a lack of normalization.
Students familiar with other RDBMS systems may be aware that some of them offer table (or its equivalent) as a storage data type. In SQL Server 2016, XML or JSON offer the equivalent functionality.

7. 20761B
Data Type Precedence
6: Working with SQL Server 2016 Data Types
Data type precedence determines which data type will be chosen when expressions of different types are combined
By default, the data type with the lower precedence is converted to the data type with the higher precedence
It is important to understand implicit conversions
Conversion to a data type of lower precedence must be made explicitly (using CAST or CONVERT functions)
Example precedence (low to high)
CHAR -> VARCHAR -> NVARCHAR -> TINYINT -> INT -> DECIMAL -> TIME -> DATE -> DATETIME2 -> XML
Not all combinations of data type have a conversion (implicit or explicit)
Note that not all the SQL Server data types are included in the precedence example—see Data Type Precedence (Transact-SQL) in the SQL Server 2016 Technical Documentation for a full list.

8. When are Data Types Converted?
20761B
When are Data Types Converted?
6: Working with SQL Server 2016 Data Types
Data type conversion scenarios
When data is moved, compared to or combined with other data
During variable assignment
Implicit conversion
When comparing data of one data type to another
Transparent to the user
Explicit conversion
Uses CAST or CONVERT functions
Code samples are fragments for illustration only.
Note that conversion functions will be covered later in the course.
Questions in workbook:
In the example, which data type will be converted? To which data type will it be converted? The char will be converted to an int.
Why does SQL Server attempt to convert the character variable to an integer and not the other way around? The data type with the lower precedence is converted to the higher.
WHERE <column of smallint type> = <value of int type>
CAST(unitprice AS INT)

9. Lesson 2: Working with Character Data
20761B
Lesson 2: Working with Character Data
6: Working with SQL Server 2016 Data Types
Demonstration: Working with Character Data
Question
You have the following query:
SELECT FirstName
FROM HumanResources.Employees
WHERE FirstName LIKE N'[^MA]%'
Will the query return an employee with the first name ‘Matthew’?
Answer
No.

10. 20761B
Character Data Types
6: Working with SQL Server 2016 Data Types
SQL Server supports two kinds of character data as fixed-width or variable-width data:
Single-byte: char and varchar
One byte stored per character
Only 256 possible characters—limits language support
Multibyte: nchar and nvarchar
Multiple bytes stored per character (usually two bytes, but sometimes up to four)
More than 65,000 characters represented—multiple language support
Precede character string literals with N (National)
text and ntext data types are deprecated, but may still be used in older systems
In new development, use varchar(max) and nvarchar(max) instead
For a detailed discussion of Unicode character support, including a discussion on Unicode characters occupying more than two bytes (the Supplementary Characters section), see Collation and Unicode Support in the SQL Server 2016 Technical Documentation.

11. Collation Collation is a collection of properties for character data
20761B
Collation
6: Working with SQL Server 2016 Data Types
Collation is a collection of properties for character data
Character set
Sort order
Case sensitivity
Accent sensitivity
When querying, collation awareness is important for comparison
Is the database case-sensitive? If so:
'Funk‘ does not equal 'funk‘
SELECT * FROM HR.Employee does not equal SELECT * FROM HR.employee
Add COLLATE clause to control collation comparison
For some students, it may be helpful to point out that all character data is stored as binary data, and that data type and collation determine how the binary data is interpreted by the database engine for comparison and display.
SELECT empid, lastname
FROM HR.employees
WHERE lastname COLLATE Latin1_General_CS_AS = N'Funk';

12. 20761B
String Concatenation
6: Working with SQL Server 2016 Data Types
The + (plus) operator and the CONCAT function can both be used to concatenate strings in SQL 2016
Using CONCAT
Converts input values to strings and converts NULL to empty string
Using + (plus)
No conversion of NULL or data type
SELECT custid, city, region, country,
CONCAT(city, ', ' + region, ', ' + country) AS location
FROM Sales.Customers
It might be helpful to students to discuss how the example CONCAT query takes advantage of the different NULL handling behavior of CONCAT and + to eliminate redundant commas in the result.
Note that the ISNULL and COALESCE functions, covered later in the course, were often previously used to convert NULLs to empty strings. CONCAT now handles that.
Note that the behavior of + when concatenating NULL values is actually dependent on the session parameter CONCAT_NULL_YIELDS_NULL (ON by default). However, setting CONCAT_NULL_YIELDS_NULL to any value other than ON is marked for deprecation, so this will become less important in future.
SELECTcustid, city, region, country,
CONCAT(city, ', ' + region, ', ' + country) AS location
FROM Sales.Customers;
SELECT empid, lastname, firstname,
firstname + N' ' + lastname AS fullname
FROM HR.Employees;

13. Character String Functions
20761B
Character String Functions
6: Working with SQL Server 2016 Data Types
Common functions that modify character strings
Function
Syntax
Remarks
SUBSTRING
SUBSTRING (expression , start , length)
Returns part of an expression.
LEFT, RIGHT
LEFT (expression , integer_value)
RIGHT (expression , integer_value)
LEFT returns left part of string up to integer_value. RIGHT returns right part of string up to integer value.
LEN, DATALENGTH
LEN (string_expression)
DATALENGTH (expression)
LEN returns the number of characters in string_expression, excluding trailing spaces. DATALENGTH returns the number of bytes used.
CHARINDEX
CHARINDEX (expressionToFind, expressionToSearch)
Searches expressionToSearch for expressionToFind and returns its start position if found.
REPLACE
REPLACE (string_expression , string_pattern , string_replacement)
Replaces all occurrences of string_pattern in string_expression with string_replacement.
UPPER, LOWER
UPPER (character_expression)
LOWER (character_expression)
UPPER converts all characters in a string to uppercase. LOWER converts all characters in a string to lowercase.
If time permits, also introduce LTRIM, RTRIM, REPLICATE and STUFF.
Note: the FORMAT function is mentioned in the workbook, though not on this slide.

14. 20761B
The LIKE Predicate
6: Working with SQL Server 2016 Data Types
The LIKE predicate can be used to check a character string for a match with a pattern
Patterns are expressed with symbols
% (Percent) represents a string of any length
_ (Underscore) represents a single character
[<List of characters>] represents a single character within the supplied list
[<Character> - <character>] represents a single character within the specified range
[^<Character list or range>] represents a single character not in the specified list or range
ESCAPE Character allows you to search for characters that would otherwise be treated as part of a pattern - %, _, [, and ])
Note that SQL Server does not natively provide support for more complex pattern-matching tools such as regular expressions.
SELECT categoryid, categoryname, description
FROM Production.Categories
WHERE description LIKE 'Sweet%';

15. Lesson 3: Working with Date and Time Data
20761B
Lesson 3: Working with Date and Time Data
6: Working with SQL Server 2016 Data Types
Demonstration: Working with Date and Time Data

16. Date and Time Data Types
20761B
Date and Time Data Types
6: Working with SQL Server 2016 Data Types
Older versions of SQL Server support only datetime and smalldatetime data types
SQL Server 2008 introduced date, time, datetime2 and datetimeoffset data types
SQL Server 2012 added further functionality for working with date and time data types
Emphasize that datetime2, date, time, and datetimeoffset are recommended for new development because they are compliant with the ANSI SQL standard, offer better accuracy, and support a wider range of dates.
The storage size for datetime2, date, and datetimeoffset can vary because these types can be defined with an optional scale parameter that limits their precision (and therefore the storage they require).
Data Type
Storage (bytes)
Date Range (Gregorian Calendar)
Accuracy
Recommended Entry Format
datetime 8
January 1, 1753 to December 31, 9999
Rounded to increments of .000, .003, or .007 seconds
YYYYMMDD hh:mm:ss[.mmm]
smalldatetime 4
January 1, 1900 to June 6, 2079
1 minute
datetime2
6 to 8
January 1, 0001 to December 31, 9999
100 nanoseconds
YYYYMMDD hh:mm:ss[.nnnnnnn]
date 3
1 day
YYYY-MM-DD
time
3 to 5
n/a – time only
hh:mm:ss[.nnnnnnn]
datetimeoffset
8 to 10
YYYY-MM-DDThh:mm:ss[.nnnnnnn][{+|-}hh:mm]

17. Entering Date and Time Data Types Using Strings
20761B
Entering Date and Time Data Types Using Strings
6: Working with SQL Server 2016 Data Types
SQL Server doesn't offer a means to enter a date or time value as a literal value
Dates and times are entered as character literals and converted explicitly or implicitly
For example, char converted to datetime due to precedence
Formats are language-dependent, and can cause confusion
Best practices:
Use character strings to express date and time values
Use language-neutral formats
SQL Server 2012 added the PARSE function to parse date string literals and indicate the culture. For example:
SELECT PARSE('7/17/2011' AS DATE USING 'en-US') AS dt;
Go to PARSE (Transact-SQL) at:
PARSE (Transact-SQL)
PARSE, TRY_PARSE and TRY_CONVERT will be covered in a later module.
Note: the range of values for the offset in datetimeoffset is +/- 14 hours (time zones). For more information, see:
datetimeoffset (Transact-SQL)
SELECT orderid, custid, empid, orderdate
FROM Sales.Orders
WHERE orderdate = ' ';

18. Working Separately with Date and Time
20761B
Working Separately with Date and Time
6: Working with SQL Server 2016 Data Types
datetime, smalldatetime, datetime2, and datetimeoffset include both date and time data
If only date is specified, time set to midnight (all zeros)
If only time is specified, date set to base date (January 1, 1900)
AS datetime2 = ' ';
AS Result;
AS time = '12:34:56';
SELECT AS datetime2) AS Result;

19. Querying Date and Time Values
20761B
Querying Date and Time Values
6: Working with SQL Server 2016 Data Types
Date values converted from character literals often omit time
Queries written with equality operator for date will match midnight
If time values are stored, queries need to account for time past midnight on a date
Use range filters instead of equality
Consider discussing the disadvantages of using a BETWEEN predicate (either missing a day or including data from an additional day) instead of >= and < as part of discussing the second example.
SELECT orderid, custid, empid, orderdate
FROM Sales.Orders
WHERE orderdate= ' ';
SELECT orderid, custid, empid, orderdate
FROM Sales.Orders
WHEREorderdate >= ' '
ANDorderdate < ' ';

20. Date and Time Functions
20761B
Date and Time Functions
6: Working with SQL Server 2016 Data Types
To get system date and time values
For example, GETDATE, GETUTCDATE, SYSDATETIME
To get date and time parts
For example, DATENAME, DATEPART
To get date and time values from their parts
For example, DATETIME2FROMPARTS, DATEFROMPARTS
To get date and time difference
For example, DATEDIFF, DATEDIFF_BIG
To modify date and time values
For example, DATEADD, EOMONTH
To validate date and time values
For example, ISDATE
Note that, for functions that construct date and time data types from parts, all arguments are required. For example, DATETIME2FROMPARTS requires eight arguments.
Not all date and time functions are available in all versions of SQL Server.
SQL Server 2016 adds the DATEDIFF_BIG function, which works in the same way as DATEDIFF but returns a bigint data type.