ACC2000: How to Rank Records Within a Query
ID: Q208946
|
The information in this article applies to:
Advanced: Requires expert coding, interoperability, and multiuser skills.
SUMMARY
This article has three examples that show you how to rank records
within a query. The first example ranks records from a table. The second
example ranks the results of a totals query. The third example demonstrates
an alternative method of handling ties in ranking.
All the examples in this article require a field that has a value that
can be evaluated using either the Greater Than (>) or the Less Than (<) operator, and a subquery to count the number of records that are greater than or less than the current record.
MORE INFORMATION
In Example 1 and Example 2 below, records following a tie in ranking are
given the same rank number they would have received had there been no tie.
The following table demonstrates a sample result with a tie in ranking:
Salesperson |
Sales |
Rank |
Peterson |
$8000 |
1 |
Wakita |
$7000 |
2 |
Akerley |
$7000 |
2 |
Reston |
$6000 |
4 |
In Example 3 below, a second query, which has a DISTINCT clause, is used to
return a unique list of ranking less than that of the current record. The
following table, using the same records as the table above, demonstrates
the results of this method:
Salesperson |
Sales |
Rank |
Peterson |
$8000 |
1 |
Wakita |
$7000 |
2 |
Akerley |
$7000 |
2 |
Reston |
$6000 |
3 |
Example 1
This example returns results that show the order in which employees were
hired.
CAUTION: Following the steps in this example will modify the sample
database Northwind.mdb. You may want to back up the Northwind.mdb file
and perform these steps on a copy of the database.
- Open the sample Northwind.mdb and create a new query based on the Employees table.
- If the property sheet is not displayed, click Properties button on the toolbar.
- Click the title bar of the Employees field list, and then set the
field list's Alias property to Emp1.
- Drag the LastName and HireDate fields from the field list to the
query grid.
- In the Sort row of the HireDate column, select Ascending.
- In the Field row of the third column, enter the following expression:
Seniority: (Select Count(*) from Employees Where [HireDate] < _
[Emp1].[HireDate];)
- Run the query. Note that the returned records are ranked sequentially
starting at 0. The subquery returns the number of employees hired before
the current employee. To rank the records starting with 1, use the
following expression in step 6 above:
Seniority: (Select Count(*) from Employees Where [HireDate] < _
[Emp1].[HireDate]+1;)
- To see the results of a tie in ranking, change the hire date of any
employee to match another (in the Employees table).
Example 2
This example returns results that rank categories by the number of products
in each category.
CAUTION: Following the steps in this example will modify the sample
database Northwind.mdb. You may want to back up the Northwind.mdb file
and perform these steps on a copy of the database.
- Create a new query based on the Products table.
- Drag the CategoryID and ProductID fields from the field list to the
query grid.
- Click the Totals button on the toolbar.
- In the Total row of the CategoryID column, select Group By. In the Total row of the ProductID column, select Count. Save the query as ProductCount.
- Create a new query based on the ProductCount query.
- Set the ProductCount field list's Alias property to Prod1.
- Drag both fields from the field list to the query grid, and then
select Descending in the Sort row of the CountofProductID column.
- In the Field row of the third column, enter the following expression.
Ranking: (Select Count(*) from ProductCount Where [CountofProductID] > _
[Prod1].[CountofProductID]) + 1
- Run the query.
Example 3
This example demonstrates an alternative method of handling ties in
ranking.
CAUTION: Following the steps in this example will modify the sample
database Northwind.mdb. You may want to back up the Northwind.mdb file
and perform these steps on a copy of the database.
- Repeat steps 1 through 8 in Example 2. After step 8, save the query as CatRanking. Run the query and note the tie in the number of products between the Seafood, Condiments, and Beverages categories.
- Create a new query based on the Products table. Drag the CategoryID
and ProductID fields from the field list to the query grid.
- Click the Totals button on the toolbar.
- In the Total row of the CategoryID column, select Group By. In the Total row of the ProductID column, select Count.
- In the Show row of the CategoryID column, clear the check box.
- Set the UniqueValues property of the query to Yes.
- Run the query. Note that the query returns a list of the six different totals of orders placed. Save the query as DistinctCount.
- View the CatRanking query in Design view. Replace the expression
in the Field row of the third column with the following expression, and
then run the query.
Ranking: (Select Count(*) from DistinctCount Where _
[CountofProductID] > Prod1.[CountofProductID]) + 1
Note that the records following a tie in ranking are given the same rank number they would have received had there been no tie.
Additional query words:
top first
Keywords : kbdta QryHowto
Version : WINDOWS:2000
Platform : WINDOWS
Issue type : kbinfo
|