Union query in excel

In the example from the previous section using the Northwind database, only data from two tables are combined. However, you can combine three or more tables very easily in a union query. For example, building on the previous example, you might want to also include the names of the employees in the query output. You can accomplish that task by adding a third query and combining with the previous SQL statement with an additional UNION keyword like this:

SELECT Customers.Company, Customers.[Last Name], Customers.[First Name]
FROM Customers

UNION

SELECT Suppliers.Company, Suppliers.[Last Name], Suppliers.[First Name]
FROM Suppliers

UNION

SELECT Employees.Company, Employees.[Last Name], Employees.[First Name]
FROM Employees

ORDER BY [Last Name], [First Name];

When you view the result in datasheet view, all employees will be listed with the sample company name, which probably isn’t very useful. If you want that field to indicate if a person is an in-house employee, from a supplier, or from a customer, you can include a fixed value in place of the company name. Here’s how the SQL would look like:

SELECT "Customer" As Employment, Customers.[Last Name], Customers.[First Name]
FROM Customers

UNION

SELECT "Supplier" As Employment, Suppliers.[Last Name], Suppliers.[First Name]
FROM Suppliers

UNION

SELECT "In-house" As Employment, Employees.[Last Name], Employees.[First Name]
FROM Employees

ORDER BY [Last Name], [First Name];

Here’s how the result appears in datasheet view. Access displays these five example records:

The query above can be reduced even further since Access only reads the names of the output fields from the first query in a union query. Here you see we’ve removed the output from the second and third query sections:

SELECT "Customer" As Employment, [Last Name], [First Name]
FROM Customers

UNION

SELECT "Supplier", [Last Name], [First Name]
FROM Suppliers

UNION

SELECT "In-house", [Last Name], [First Name]
FROM Employees

ORDER BY [Last Name], [First Name];

In an Access union query, ordering is allowed only once but each query can be filtered individually. Building on the previous section’s union query, here’s an example of where we’ve filtered each query by adding a WHERE clause.

SELECT "Customer" As Employment, Customers.[Last Name], Customers.[First Name]
FROM Customers
WHERE [State/Province] = "UT"

UNION

SELECT "Supplier", [Last Name], [First Name]
FROM Suppliers
WHERE [Job Title] = "Sales Manager"

UNION

SELECT "In-house", Employees.[Last Name], Employees.[First Name]
FROM Employees
WHERE City = "Seattle"

ORDER BY [Last Name], [First Name];

Switch to datasheet view and you’ll see results similar to this:

If the queries to union are very different, you might encounter a situation where an output field must combine data of different data types. If so, the union query most often will return the results as a text data type since that data type can hold both text and numbers.

To understand how this works, we’ll use the Product Transactions union query in the Northwind sample database. Open that sample database and then open the Product Transactions query in datasheet view. The last ten records should be similar to this output:

Let’s assume you want the Quantity field split into two — Buy and Sell. Let’s also assume you want to have a fixed zero value for the field with no value. Here’s how the SQL will look for this union query:

SELECT [Product ID], [Order Date], [Company Name], [Transaction], 0 As Buy, [Quantity] As Sell
FROM [Product Orders]

UNION

SELECT [Product ID], [Creation Date], [Company Name], [Transaction], [Quantity] As Buy, 0 As Sell
FROM [Product Purchases]

ORDER BY [Order Date] DESC; 

If you switch to datasheet view, you’ll see the last ten records now displayed like the following:

Continuing this example, what if you want the fields with zero to be empty? You can modify the SQL to display nothing instead of zero by adding in the Null keyword like the following:

SELECT [Product ID], [Order Date], [Company Name], [Transaction], Null As Buy, [Quantity] As Sell
FROM [Product Orders]

UNION

SELECT [Product ID], [Creation Date], [Company Name], [Transaction], [Quantity] As Buy, Null As Sell
FROM [Product Purchases]

ORDER BY [Order Date] DESC;

However, as you might have observed switching to datasheet view, you now have an unexpected result. In the Buy column, every field is cleared:

The reason this happens is because Access determines the data types of the fields from the first query. In this example, Null is not a number.

So what happens if you try and insert an empty string for the blank value of fields? The SQL for this attempt might look like the following:

SELECT [Product ID], [Order Date], [Company Name], [Transaction], "" As Buy, [Quantity] As Sell
FROM [Product Orders]

UNION

SELECT [Product ID], [Creation Date], [Company Name], [Transaction], [Quantity] As Buy, "" As Sell
FROM [Product Purchases]

ORDER BY [Order Date] DESC;

When you switch to datasheet view, you’ll see that Access retrieves the Buy values, but it converted the values to text. You can tell these are text values since they are left-aligned in the datasheet view. The empty string in the first query is not a number which is why you see these results. You’ll also notice that the Sell values are also converted to text because the purchase records contain an empty string.

So how do you solve this puzzle?

A solution is to force the query to expect the field value to be a number. That can be accomplished with the expression:

IIf(False, 0, Null)

The condition to check, False, will never be True, thus the expression will always return Null, but Access still evaluates both output options and decides the output to be numeric or Null.

Here’s how we can use this expression in our working example:

SELECT [Product ID], [Order Date], [Company Name], [Transaction], IIf(False, 0, Null) As Buy, [Quantity] As Sell
FROM [Product Orders]

UNION

SELECT [Product ID], [Creation Date], [Company Name], [Transaction], [Quantity] As Buy, Null As Sell
FROM [Product Purchases]

ORDER BY [Order Date] DESC;

Note that it is not necessary to modify the second query.

If you switch to datasheet view, you’ll now see a result that we want:

An alternative method to achieve the same result is to prepend the queries in the union query with yet another query:

SELECT 
    0 As [Product ID], Date() As [Order Date], 
    "" As [Company Name], "" As [Transaction], 
    0 As Buy, 0 As Sell
FROM [Product Orders]
WHERE False

For each field, Access returns fixed values of the data type you define. Of course, you don’t want the output of this query to interfere with the results so the trick to avoid that is to include a WHERE clause to False:

WHERE False

This is a little trick since this is always false and then the query doesn’t return anything. Combining this statement with the existing SQL and we arrive at a completed statement as follows:

SELECT 
    0 As [Product ID], Date() As [Order Date], 
    "" As [Company Name], "" As [Transaction], 
    0 As Buy, 0 As Sell
FROM [Product Orders]
WHERE False

UNION

SELECT [Product ID], [Order Date], [Company Name], [Transaction], Null As Buy, [Quantity] As Sell
FROM [Product Orders]

UNION

SELECT [Product ID], [Creation Date], [Company Name], [Transaction], [Quantity] As Buy, Null As Sell
FROM [Product Purchases]

ORDER BY [Order Date] DESC;

Note: The combined query here in this example using the Northwind database returns 100 records, while the two individual queries return 58 and 43 records for a total of 101 records. The reason for this discrepancy is because two records are not unique. See the section, Working with distinct records in union queries using UNION ALL, to learn how to solve this scenario by using UNION ALL.

A special case for a union query is to combine a set of records with one record that contains the sum of one or more fields.

Here is another example that you can create in the Northwind sample database to illustrate how to get a total in a union query.

1. Create a new simple query to view the purchase of beers (Product ID=34 in the Northwind database) using the following SQL syntax:

   SELECT [Purchase Order Details].[Date Received], [Purchase Order Details].Quantity
   FROM [Purchase Order Details]
   WHERE ((([Purchase Order Details].[Product ID])=34))
   
   ORDER BY [Purchase Order Details].[Date Received];

2. Switch to datasheet view and you should see four purchases:

   Date Received	Quantity
   1/22/2006	100
   1/22/2006	60
   4/4/2006	50
   4/5/2006	300

3. To obtain the total, create a simple aggregating query using the following SQL:

   SELECT Max([Date Received]), Sum([Quantity]) AS SumOfQuantity
   FROM [Purchase Order Details]
   WHERE ((([Purchase Order Details].[Product ID])=34))

4. Switch to datasheet view and you should see only one record:

   MaxOfDate Received	SumOfQuantity
   4/5/2006	510

5. Combine these two queries into a union query to append the record with the total quantity to the purchase records:

   SELECT [Purchase Order Details].[Date Received], [Purchase Order Details].Quantity
   FROM [Purchase Order Details]
   WHERE ((([Purchase Order Details].[Product ID])=34))
   
   UNION
   
   SELECT Max([Date Received]), Sum([Quantity]) AS SumOfQuantity
   FROM [Purchase Order Details]
   WHERE ((([Purchase Order Details].[Product ID])=34))
   
   ORDER BY [Purchase Order Details].[Date Received];

6. Switch to datasheet view and you should see the four purchases with the sum of each followed by a record that totals the quantity:

   Date Received	Quantity
   1/22/2006	60
   1/22/2006	100
   4/4/2006	50
   4/5/2006	300
   4/5/2006	510

That covers the basics of adding totals into a union query. You might also want to include fixed values in both queries such as “Detail” and “Total” to visually separate the total record from the other records. You can review using fixed values in the section Combine three or more tables or queries in a union query.

Union queries in Access by default only include distinct records. But what if you want to include all records? Another example might be useful here.

In the previous section, we showed you how to create a total in a union query. Modify that union query SQL to include Product ID= 48:

SELECT [Purchase Order Details].[Date Received], [Purchase Order Details].Quantity
FROM [Purchase Order Details]
WHERE ((([Purchase Order Details].[Product ID])=48))

UNION

SELECT Max([Date Received]), Sum([Quantity]) AS SumOfQuantity
FROM [Purchase Order Details]
WHERE ((([Purchase Order Details].[Product ID])=48))

ORDER BY [Purchase Order Details].[Date Received];

Switch to datasheet view and you should see a somewhat misleading result:

One record of course doesn’t return twice the quantity in total.

The reason you see this result is because on one day the same quantity of chocolates was sold twice — as recorded in the Purchase Order Details table. Here’s a simple select query result showing both records in the Northwind sample database:

In the union query previously noted, you can see that the Purchase Order ID field isn’t included and that the two fields don’t make up two distinct records.

If you want to include all records, use UNION ALL instead of UNION in your SQL. This will mostly likely have an impact on the sorting of the results, so you might want to also include an ORDER BY clause to determine a sort order. Here’s the modified SQL building off the previous example:

SELECT [Purchase Order Details].[Date Received], Null As [Total], [Purchase Order Details].Quantity
FROM [Purchase Order Details]
WHERE ((([Purchase Order Details].[Product ID])=48))

UNION ALL

SELECT Max([Date Received]), "Total" As [Total], Sum([Quantity]) AS SumOfQuantity
FROM [Purchase Order Details]
WHERE ((([Purchase Order Details].[Product ID])=48))

ORDER BY [Total];

Switch to datasheet view and you should see all the details in addition to a total as the last record:

A common usage for a union query is to serve as the record source for a combo box control on a form. You can use that combo box to select a value to filter the form’s records. For example, filtering the employee records by their city.

To see how this might work, here’s another example that you can create in the Northwind sample database to illustrate this scenario.

1. Create a simple select query using this SQL syntax:

   SELECT Employees.City, Employees.City AS Filter
   FROM Employees;

2. Switch to datasheet view and you should see the following results:

   City	Filter
   Seattle	Seattle
   Bellevue	Bellevue
   Redmond	Redmond
   Kirkland	Kirkland
   Seattle	Seattle
   Redmond	Redmond
   Seattle	Seattle
   Redmond	Redmond
   Seattle	Seattle

3. Looking at those results you might not see a lot of value. Expand the query though and transform it to a union query by using the following SQL:

   SELECT Employees.City, Employees.City AS Filter
   FROM Employees
   
   UNION
   
   SELECT "<All>", "*" AS Filter
   FROM Employees
   
   ORDER BY City;

4. Switch to datasheet view and you should see the following results:

   City	Filter
   <All>	*
   Bellevue	Bellevue
   Kirkland	Kirkland
   Redmond	Redmond
   Seattle	Seattle

   Access performs a union of the nine records, previously shown, with fixed field values of <All> and «*».

   Since this union clause doesn’t contain UNION ALL, Access returns only distinct records which means each city is returned only once with fixed identical values.

5. Now that you have a completed union query displaying each city name only once, along with an option that effectively selects all cities, you can use this query as the record source for a combo box on a form. Using this specific example as a model, you could create a combo box control on a form, set this query as its record source, set the Column Width property of the Filter column to 0 (zero) to hide it visually, and then set the Bound Column property to 1 to indicate the index of the second column. In the Filter property of the form itself, you can then add in code such as the following to activate a form filter using the value of what was selected in the combo box control:

   Me.Filter = "[City] Like '" & Me![FilterComboBoxName].Value & "'"
   Me.FilterOn = True

   The user of the form can then filter the form records to a specific city name or select <All> to list all records for all cities.

Top of Page