We have used the CoalesceEmpty() function, in its numeric "incarnation," to replace nulls with a number, zero. The string version of CoalesceEmpty() works in much the same manner, except that it substitutes "string for string."
Let's take a look at an example where we again use CoalesceEmpty(), but where the requirement includes a bit of a twist: We will be asked to replace nulls occurring within a numeric value with a string a word versus a number. We will base the scenario upon the original core query, taking the following setting into consideration.
Let's say that we have discussed the results in our last dataset with the Logistics group, which is happy with the fact that we are able to make substitutions for the nulls, but wishes to make a change. The consumers have determined that the presentation will be enhanced by the use of a word, "None," instead of zeros (apparently this was suggested by an overly excitable senior manager) where no shipping volume of a given product occurred for the year. We agree to make the change, and proceed as follows:
15. Within the query we have saved as MDX033-2, replace the existing comment line of the query with the following:
-- MDX033-3, Use COALESCEEMPTY()
to Replace a Null Numeric Val Exp w String
16. Save the query as MDX033-3 to protect its predecessor.
17. Replace the syntax between the above comment line and the SELECT keyword in the existing query, currently appearing as:
'COALESCEEMPTY( [Measures].[Units Shipped], 0)'
With the following:
' "None" '
'CoalesceEmpty([Measures].[Units Shipped], [Measures].[NoneShipped])'
We are replacing the calculated member we used initially with two calculated members, the second of which relies upon the first to store the string that we wish to return, under the cloak of another measure, to be substituted for null values that are encountered. Because we have assigned the output of the second calculated member to the name we have used within our SELECT statement [Measures].[Vol Shipped]), further alterations within the query are not necessary.
The Query pane appears as shown in Illustration 4, with the new additions highlighted.
Illustration 4: CoalesceEmpty() Arranged to Substitute Null Numeric Value with a String
18. Execute the query by clicking the Run Query button in the toolbar.
The Results pane is populated, and the dataset partially depicted in Illustration 5 appears.
Illustration 5: Partial Result Dataset String Substituted for Null Numeric Value
We note that the nulls have been replaced by the word "None," as requested by the information consumers.
19. Re-save the file as MDX033-3.
We deliver the new results to the Logistics information consumers, who accept with thanks the results as meeting the current need.
20. Select File --> Exit to close the MDX Sample Application, when ready.
In this lesson, we explored the MDX CoalesceEmpty() function. Beginning with a discussion of the pervasiveness of sparsity within OLAP cubes as a circumstance inherent to their nature, we introduced CoalesceEmpty() as an avenue to handling that sparsity from a presentation perspective. We commented upon the general operation of CoalesceEmpty(), and then we examined the syntax of the function.
We undertook illustrative examples to gain hands-on experience with the function in practice exercises, where we described a business need as specified by a hypothetical group of information consumers. We examined a simple case where we used CoalesceEmpty() to substitute a numeric value for a null numeric value. We then considered a more complicated scenario, where we went beyond the standard uses (substitution of numeric value for numeric value, or string value for string value) and substituted a string value for a null numeric value through the use of stepped calculated members. Finally, we briefly discussed the results datasets we obtain in the practice examples.
» See All Articles by Columnist William E. Pearson, III
Discuss this article in the MSSQL Server 2000 Analysis Services and MDX Topics Forum.