Using DataSet Attributes for Sorting

Pattern datasets identify the datasets that you want to use in your pattern. They allow you to specify how to sort your datasets. Using the Edit Tag function to define sort order, use the following guidelines:

• You can specify multiple attributes (comma-separated) when defining the sorting sequence.
• You can intermingle multiple datasets based on the sorting attributes; the dataset with the few- est sorting attributes drives the extent of inter-dataset mixing. The sorting algorithm uses the lowest common denominator approach. For example, if you have two datasets with the following attributes and the following sort orders:

  DATASET1: INT1, FLOAT1, STRING1 SORT ORDER: INT1, FLOAT1, STRING1 DATASET2: INT2, FLOAT2, STRING2 SORT ORDER: INT2, FLOAT2

  The sorting will only sort by INT1/INT2 and FLOAT1/FLOAT2 leaving STRING1/STING2 since, a similar attribute does not exist in DATASET2 defined sort order.

• Use the same data type for sorting attributes in each dataset.
• Arrange the datasets so that the system handles tie-breaks correctly; invert the list if you want to run the job in reverse. In cases where a sorting attribute has the same value in two datasets, the order of retrieval is based on the dataset order.