Is collecting stats on (PARTITION, PI) or (PARTITION, PI, partitioning column), subject to the 16 byte limit of multi-column stats? That is, if a table has a PI that is on a really long set of column(s), will the stats on PARTITION+PI be misleading to the optimizer if the first 16 bytes aren't distinct enough?
Hello Carrie,
Is collecting stats on (PARTITION, PI) or (PARTITION, PI, partitioning column), subject to the 16 byte limit of multi-column stats? That is, if a table has a PI that is on a really long set of column(s), will the stats on PARTITION+PI be misleading to the optimizer if the first 16 bytes aren't distinct enough?
Thanks!