Restrict varchar() column to specific values?

Question

Is there a way to specify, for example 4 distinct values for a varchar column in MS SQL Server 2008?

For example, I need a column called Frequency (varchar) that only accepts 'Daily', 'Weekly', 'Monthly', 'Yearly' as possible values

Is this possible to set within the SQL Server Management Studio when creating the table?

Answer 1

You want a check constraint.

CHECK constraints determine the valid values from a logical expression that is not based on data in another column. For example, the range of values for a salary column can be limited by creating a CHECK constraint that allows for only data that ranges from $15,000 through $100,000. This prevents salaries from being entered beyond the regular salary range.

You want something like:

ALTER TABLE dbo.Table ADD CONSTRAINT CK_Table_Frequency
    CHECK (Frequency IN ('Daily', 'Weekly', 'Monthly', 'Yearly'))

You can also implement check constraints with scalar functions, as described in the link above, which is how I prefer to do it.

Answer 2

Have you already looked at adding a check constraint on that column which would restrict values. Something like below:-

CREATE TABLE SomeTable
(
Id int NOT NULL,
Frequency varchar(200)
CONSTRAINT chk_Frequency CHECK Frequency IN ('Daily', 'Weekly', 'Monthly', 'Yearly')
)

Answer 3

When you are editing a table
Right Click -> Check Constraints -> Add -> Type something like Frequency IN ('Daily', 'Weekly', 'Monthly', 'Yearly') in expression field and a good constraint name in (Name) field.
You are done.

Answer 4

Personally, I'd code it as tinyint and:

• Either: change it to text on the client, check constraint between 1 and 4
• Or: use a lookup table with a foreign key

Reasons:

• It will take on average 8 bytes to store text, 1 byte for tinyint. Over millions of rows, this will make a difference.

• What about collation? Is "Daily" the same as "DAILY"? It takes resources to do this kind of comparison.

• Finally, what if you want to add "Biweekly" or "Hourly"? This requires a schema change when you could just add new rows to a lookup table.