The number of distinct values is more significant for these types of multicolumn statistics you are asking about below, than is the length of the values stored in the statistics histogram.
When you issue a COLLECT STATISTICS statement, the aggregation activity calculates the number of distinct values before any truncation is performed, so that metric is a more reliable value than would be an interval value truncated at 16 bytes.
The number of distinct values is more significant for these types of multicolumn statistics you are asking about below, than is the length of the values stored in the statistics histogram.
When you issue a COLLECT STATISTICS statement, the aggregation activity calculates the number of distinct values before any truncation is performed, so that metric is a more reliable value than would be an interval value truncated at 16 bytes.
Thanks, -Carrie