Sergei Golubchik added a comment - 2016-05-17 07:53 What would it be really useful for?

Anders Karlsson added a comment - 2016-05-17 08:50 This is useful in the case where there are a lot of partitions and a lot of data. Partitioning is useful in providing ease and performance of data pruning, in the sense that old data can be removed by dropping partitions, as well as a means to increased performance in that the amount of data to process can be kept smaller and partitions can be pruned. When there is a lot of data, these two use cases, partition dropping and performance boost, are combined using subpartitioning, where the main partition is used to allow partitions with outdated data to be dropped and the subpartitions allow data to be split in smaller chunks for performance reasons and partition pruning. Now, current HASH and KEY partitioning schemes are rather basic and combining them with LIST and RANGE partitioning would ease subpartitioning to allow the requested partition pruning to be more effective and useful.