hi ,
We have a large volume partitioned table. Each partition running into 5-10 million rows. We daily load into this table and perform join queries immediately with table against other tables. In the interest of time we copy stats from previous partition everyday before we load and at the end of the day we collect stats . Recently we perfomed DDL against this table (added a couple of extra columns to the table) and we copied stats as usual from older partitions and started seeing very funning execution plans (sub optimal nested loop plans).
The question is when we perform DDL against any table is it necessary to gather stats of the entire table ? Does stats get stale or unusable because of DML. Also when we copy stats from partitions which were created a populated before the DML is performed does it causes issue in the optimizer.