What is the Difference between ROW_NUMBER(), RANK() and DENSE_RANK()?

Introduction

All of these three functions are used to calculate RowID for the result set returned from a query but in a slightly different way.

Row_Number()

This function will assign a unique id to each row returned from the query.

Consider the following query:

DECLARE @Table TABLE (
      Col_Value varchar(2)
)

INSERT INTO @Table (Col_Value)
      VALUES ('A'),('A'),('A'),('B'),('B'),('C'),('C');

SELECT
      Col_Value,
      ROW_NUMBER() OVER (ORDER BY Col_Value) AS 'RowID'
FROM
      @Table;    

After executing it, we will get:

As we notice, each and every row has a unique ID.

Rank()

This function will assign a unique number to each distinct row, but it leaves a gap between the groups. Let me explain with a query, we will use the same query we used above with Rank().

SELECT
      Col_Value,
      Rank() OVER (ORDER BY Col_Value) AS 'RowID'
FROM
      @Table;

This query will return:

As we can see, rowid is unique for each distinct value, but with a gap. What is this gap?

This gap represents number of occurrence. For example: value ‘a’ is repeated thrice and has rank ‘1’, the next rank will be 1+3=4. Same with the next value 4+2=6 and so on.

Dense_Rank()

This function is similar to Rank with only difference, this will not leave gaps between groups.

So if we use the same query used for Rank, then:

SELECT
      Col_Value,
      DENSE_RANK() OVER (ORDER BY Col_Value) AS 'RowID'
FROM
      @Table;

We will get:

So it is clear that it generates a unique id for each group and without repetition.

As we are clear now what these functions do, we can use them in different scenarios based on our need.

For example: Row_Number() can be used if we want to skip and fetch some records. Take a look at this query:

WITH AllRecords
AS (
      SELECT
            Col_Value,
            ROW_NUMBER() OVER (ORDER BY Col_Value) AS 'RowID'
      FROM
            @Table)

SELECT
      Col_Value,
      RowID
FROM
      AllRecords
WHERE
      RowID >= 4
      AND RowID <= 5;

This will return only those rows with RowID 4 and 5. This function is very useful in paging data in SQL instead of in code till SQL SERVER 2012. In SQL Server 2012, Microsoft introduces a new feature Offset Fetch similar to what we did in the above query.

In SQL Server 2012, the above query will be written as:

SELECT
      Col_Value
FROM
      @Table
ORDER BY
      Col_Value OFFSET 3 ROWS FETCH NEXT 2 ROWS ONLY;

This will skip first 3 rows and will fetch next 2.

Similarly, we can use Dense_Rank() when we need to calculate rowid with Select Distinct.

SELECT
DISTINCT
      Col_Value,
      DENSE_RANK() OVER (ORDER BY Col_Value) AS 'RowID'
FROM
      @Table;

Will return:-

Or when we need a report where ranks clashes. For example: Two students scored same marks and share same division.

We can use Rank() if we don’t want consecutive numbers.

Hope this tip helped you.