Appropriate datatype for holding percent values?

What is the best datatype for holding percent values ranging from 0.00% to 100.00%?

Assuming two decimal places on your percentages, the data type you use depends on how you plan to store your percentages:

• If you are going to store their fractional equivalent (e.g. 100.00% stored as 1.0000), I would store the data in a decimal(5,4) data type with a CHECK constraint that ensures that the values never exceed 1.0000 (assuming that is the cap) and never go below 0 (assuming that is the floor).
• If you are going to store their face value (e.g. 100.00% is stored as 100.00), then you should use decimal(5,2) with an appropriate CHECK constraint.

Combined with a good column name, it makes it clear to other developers what the data is and how the data is stored in the column.

• Hold as a decimal.
• Add check constraints if you want to limit the range (e.g. between 0 to 100%; in some cases there may be valid reasons to go beyond 100% or potentially even into the negatives).
• Treat value 1 as 100%, 0.5 as 50%, etc. This will allow any math operations to function as expected (i.e. as opposed to using value 100 as 100%).
• Amend precision and scale as required (these are the two values in brackets columnName decimal(precision, scale). Precision says the total number of digits that can be held in the number, scale says how many of those are after the decimal place, so decimal(3,2) is a number which can be represented as #.##; decimal(5,3) would be ##.###.
• decimal and numeric are essentially the same thing. However decimal is ANSI compliant, so always use that unless told otherwise (e.g. by your company's coding standards).

Example Scenarios

• For your case (0.00% to 100.00%) you'd want decimal(5,4).
• For the most common case (0% to 100%) you'd want decimal(3,2).
• In both of the above, the check constraints would be the same

Example:

if object_id('Demo') is null
create table Demo
	(
		Id bigint not null identity(1,1) constraint pk_Demo primary key
		, Name nvarchar(256) not null constraint uk_Demo unique 
		, SomePercentValue decimal(3,2) constraint chk_Demo_SomePercentValue check (SomePercentValue between 0 and 1)
		, SomePrecisionPercentValue decimal(5,2) constraint chk_Demo_SomePrecisionPercentValue check (SomePrecisionPercentValue between 0 and 1)
	)

Further Reading:

• Decimal Scale & Precision: http://msdn.microsoft.com/en-us/library/aa258832%28SQL.80%29.aspx
• 0 to 1 vs 0 to 100: https://stackoverflow.com/questions/3304177/c-storing-percentages-50-or-0-50
• Decimal vs Numeric: https://stackoverflow.com/questions/759401/is-there-any-difference-between-decimal-and-numeric-in-sql-server

I agree with Thomas and I would choose the DECIMAL(5,4) solution at least for WPF applications.

Have a look to the MSDN Numeric Format String to know why : http://msdn.microsoft.com/en-us/library/dwhawy9k#PFormatString

> The percent ("P") format specifier multiplies a number by 100 and converts it to a string that represents a percentage.

Then you would be able to use this in your XAML code:

DataFormatString="{}{0:P}"

If 2 decimal places is your level of precision, then a "smallint" would handle this in the smallest space (2-bytes). You store the percent multiplied by 100.

EDIT: The decimal type is probably a better match. Then you don't need to manually scale. It takes 5 bytes per value.

Use numeric(n,n) where n has enough resolution to round to 1.00. For instance:

declare @discount numeric(9,9)
	, @quantity int
select @discount = 0.999999999
	, @quantity = 10000

select convert(money, @discount * @quantity)