The ROW list is used when either the sparse or the dense format of the CONDATA= data set is being used. SAS variables in the ROW list have values that are constraint or special row names. The SAS variables in the ROW list must have character values.

If the dense data format is used, there must be only one SAS variable in this list. In this case, if a ROW list is not specified, the CONDATA= data set is searched and the SAS variable with the name _ROW_ or _CON_ is used. If that search fails to find a suitable SAS variable, data for each constraint must reside in only one observation.

If the sparse data format is used and the ROW statement is not specified, the CONDATA= data set is searched and SAS variables with names beginning with _ROW or _CON are used. The number of SAS variables in the ROW list must not be less than the number of SAS variables in the COEF list. The th ROW list variable is paired with the th COEF list variable. If the number of ROW list variables is greater than the number of COEF list variables, the last ROW list variables have no COEF partner. These ROW list variables that have no corresponding COEF list variable are used in observations that have a TYPE list variable value. All ROW list variable values are tagged as having the type indicated. If there is no TYPE list variable, all ROW list variable values are constraint names.