Database Journal
MS SQL Oracle DB2 Access MySQL PostgreSQL Sybase PHP SQL Etc SQL Scripts & Samples Tips Database Forum Rss Feed

» Database Journal Home
» Database Articles
» Database Tutorials
MS Access
SQL Scripts & Samples
» Database Forum
» Slideshows
Free Newsletters:

News Via RSS Feed

Rss Feed

Database Journal |DBA Support |SQLCourse |SQLCourse2

Featured Database Articles


Posted Mar 7, 2005

MDX Essentials: Set and String Functions: The GENERATE() Function - Page 2

By William Pearson

The GENERATE() Function


The GENERATE() function, according to the Analysis Services Books Online, "applies a set to each member of another set and joins the resulting sets by union." The Books Online goes on to say that GENERATE() "alternatively, returns a concatenated string created by evaluating a string expression over a set." We will examine the way the function accomplishes these combinations, eliminating duplicates automatically (while allowing us the option to include them, if necessary), in the sections that follow.

We will examine the syntax for the GENERATE() function in general, building to operations upon sets in practice exercises, within which we will meet a hypothetical business need. In this way, we will be able to clearly see that the GENERATE() function does, in fact, produce the results we might expect. Our objective is to gain a richer understanding of the capabilities found within the GENERATE() function, together with a feel for its similarities to the CROSSJOIN() function, which we have discussed in Basic Set Functions: The CROSSJOIN() Function, as well as our previous article, The CROSSJOIN() Function: Breaking Bottlenecks.


The GENERATE() function comes in two "models:" The Books Online refer to these as a "set version" and a "string version." In the set version, the function generates a results set based upon the application of a specified secondary set (which often itself contains a function) to a specified primary set. As we shall see, the function conducts itself much like the CROSSJOIN() function in many cases - most notably when the secondary set is composed of a more or less fixed group of members. We will also see that the power of GENERATE() is leveraged significantly when we go beyond a relatively fixed set of members in the secondary set, and construct the secondary set via an expression that specifies the primary set's current member.

The string version provides for the concatenation of a string expression (substituted in the place of the secondary set appearing in the set version) with each element appearing in the primary set. A delimiter can be specified to separate the elements, as well, if this is useful to the end result (as it proves to be in a practice example we undertake in a later section).

Let's look at some syntax illustrations to further clarify the operation of GENERATE().


The set version of GENERATE() resembles closely the string version, with regard to syntax. In the former, the primary and secondary sets upon which the operation of the function is to be performed are placed within the parentheses to the right of GENERATE. The set version applies Set2 to each member of Set1, performing a union of the resulting sets. We can direct that duplicates in the results are retained by specifying ALL, but the default behavior is to eliminate duplicates. The syntax is shown in the following string:

GENERATE( <<Set1>>, <<Set2>>[, ALL] )

The string version of GENERATE() appears as follows:

GENERATE( <<Set>>, <<String Expression>>[, <<Delimiter>>])

Iterating through each member of the set specified in <<Set>> above, this version of the function evaluates the specified <<String Expression>> against the respective member and returns a concatenation between the two in each case. The member and the evaluated <<String Expression>> can be delimited in the return string with the optionally supplied <<Delimiter>>, should we desire separation of the two components in the concatenated string that is returned.

The following simple example illustrates conceptually the operation of the GENERATE() function, set version (by far the more useful and pervasive version). It also shows that, within the context of simpler requirements, we can often obtain the same results with a seemingly less complex approach. The example then illustrates a more elaborate scenario, where the GENERATE() approach is certainly more efficient.

NOTE: We will be doing a practice exercise in subsequent sections, but if you want to "test drive" the below samples, the syntax will work if it is cut and pasted, or typed, into the MDX Sample Application. I have often found "fragments" in discussions such as this less than useful, when one is trying to learn new techniques, and so forth. The fact that it is easier for the author makes the practice commonplace, but it is one of many aggravating aspects of technical publishing that I hope to continue to avoid).

Let's say we have a requirement to return the top three cities in the states of California and Washington with regard to Units Shipped, one of several measures stored within the sample FoodMart Warehouse cube that accompanies an Analysis Services installation. We can achieve our objectives by employing the set version of the GENERATE() function as follows:

   {[Measures].[Units Shipped]} ON COLUMNS,
      ({[Store].[All Stores].[USA].[CA], 
        [Store].[All Stores].[USA].[WA]},
              [Store].[Store City]),
                 3, [Measures].[Units Shipped]))} ON ROWS

The query results would appear as depicted in Table 1.


Units Shipped

Los Angeles


San Diego


Beverly Hills








Table 1: Results of the GENERATE() Function, Selecting Units Shipped as the Measure

We can obtain identical results with the following query:

   {[Measures].[Units Shipped]} ON COLUMNS,
      {[Store].[All Stores].[USA].[CA].Children}, 
          3, [Measures].[Units Shipped]),
                {[Store].[All Stores].[USA].[WA].Children}, 
                   3, [Measures].[Units Shipped])} ON ROWS

The second query may seem more intuitive to many of us, and certainly presents indirect insight into the operation of the GENERATE() function. Intuitive or not, however, the GENERATE() function can certainly be the compact alternative in more elaborate uses. Consider the following query:


   {[Measures].[Units Shipped]} ON COLUMNS,




            [Warehouse].[Warehouse Name]),

               1, [Measures].[Units Shipped]))} ON ROWS





The query results would appear as shown in Table 2.


Units Shipped

Bellmont Distributing


Rose Food Warehousing


Freeman And Co.


Derby and Hunt


Salka Warehousing


Focus, Inc.


Jamison, Inc.


Bastani and Sons


Anderson Warehousing


Worthington Food Products


Big Quality Warehouse


Artesia Warehousing, Inc.


Jorgensen Service Storage


Food Service Storage, Inc.


Quality Distribution, Inc.


Treehouse Distribution


Foster Products


Destination, Inc.


Quality Warehousing and Trucking


Jones International


Jose Garcia, Inc.


Valdez Warehousing


Maddock Stored Foods


Table 2: Results of the Second Query Example Containing the GENERATE() Function

The above example, where our query is retrieving the Warehouse in each individual Warehouse City, together with the largest quantity of Units Shipped, represents a scenario where the GENERATE() approach is more concise than alternative approaches. To achieve the same result, we would be forced to employ TOPCOUNT() for each Warehouse City present in the Warehouse cube. This would be cumbersome, at best, and result in a far lengthier query than the compact query we can achieve using GENERATE().

We will activate the concepts involved in the foregoing discussions by practicing the use of the GENERATE() function in the section that follows. As part of our practice, we will undertake examples with each of the set and string versions of the function.

MS SQL Archives

Latest Forum Threads
MS SQL Forum
Topic By Replies Updated
SQL 2005: SSIS: Error using SQL Server credentials poverty 3 August 17th, 07:43 AM
Need help changing table contents nkawtg 1 August 17th, 03:02 AM
SQL Server Memory confifuration bhosalenarayan 2 August 14th, 05:33 AM
SQL Server Primary Key and a Unique Key katty.jonh 2 July 25th, 10:36 AM