With Teradata 14.10, we now have the ability to collect stats not only on columns, but on tables and expressions as well. With better analysis and consideration, you can achieve good performance improvement in queries.
The way you implement collecting statistics plays a vital role in improving performance. We need to discover the optimal way to make use of statistics. Collecting statistics blindly on columns or expressions may help one time, but in the long term it may degrade the performance of the query.
Once we identify the column on which we are ready to collect statistics, we need to identify the type of stats to be collected. Should we collect on them individually, or combine them all into one?
