Hi @Uma Maheswara Rao Desula​,

This is the same link you shared previously. This article says about inferring partition predicate from a joined dictionary table. In such a case the predicate is not mentioned in the query, but it can inferred according to the query logic (this is why it is called dynamic). The optimizer understands that the JOIN predicate is equal (or a subset) of partitioning predicate, so it can utilize partitioning. As the first step it filters the dictionary table to find the values that contribute to JOIN. When done, it used partition pruning based on values from the first step. Exact partitions cannot be identified prior to execution (those depend on the content in the dictionary table), so the dynamic style. Same partitioning technology, but a bit smarter.

In my case what could be the logic for the engine that it can use to utilize partitioning? "Time" column is not in partitioning predicate, and "day" column is filled manually, so the engine does not have information to convert predicate for "Time" into predicate for "day".