When you set the criteria in a decision-campaign Filter node or a calculated variable, it is possible to select a LIKE operator.

The LIKE operator selects records by comparing character strings with a pattern-matching specification. The percent sign (%) is used as a wildcard and matches any sequence of zero or more characters. The underscore character (_) is also used as a wildcard and matches a single character.

Therefore, you can create criteria such as the following (this is a Filter node example):

There is, however, a specific set of circumstances where the match incorrectly fails and the record is not selected. Those circumstances are met when all of the following are true: 

• The pattern-matching specification includes the % sign.
• The % sign is followed by one or more characters (as shown in the example above).
• There is a partial match followed by a complete match.

So, in the example above, if the campaign name has a value of CXAB, then the record is selected as expected.

But if the campaign name has a value of AXAB, then the match fails and the record is not selected because there is a partial match with the first A in the campaign name.  

The workaround for this problem is to use the INDEX function instead of the LIKE operator, as follows:

1. In your Filter node properties, click the Select data item icon. Then click the plus sign (+) sign to create a calculated variable.
2. In the Calculated Variable Properties window:
   1. Enter a calculated variable name and description.
   2. Choose Boolean as the type.
   3. Insert this expression:
      INDEX(target-expression,search-expression)

      In this expression:

      • target-expression is the name of the data item that you want to filter on
      • search-expression is the character string that must be found in the data item in order for the record to pass the filter (for example, CXAB)

      The following expression is an example that shows values for target-expression and search-expression:

      INDEX(<<campaign.name>>, 'CXAB')
   4. Click OK. Then select the newly created calculated variable and click OK again.
3. In the Filter node properties, click the Select value icon in the Values column and select Yes.
4. In Filter node properties, click OK again to confirm the changes that you made in the Filter node.

There is a further but similar issue with a combination of the underscore (_) and the percent (%) wildcard characters in the pattern-matching specification.

Consider this example:

It is expected that if the campaign name has a value of 2000, then the match will be successful and the record will be selected. However, this match fails because the pattern-matching specification contains a percent sign (%) followed by an underscore (_). In this case (and in all similar cases—for example, 2123, 2ABC, and so on), the match incorrectly fails.   

Click the Hot Fix tab in this note to access the hot fix for this issue.

Operating System and Release Information

A fix for this issue for SAS Decision Services 6.4_M1 is available at:

A fix for this issue for SAS Decision Services 6.4 is available at: