Columnstore - Doubts

Help
1

  1. Is there any way to query the aggregator node metadata? and only data from a single node?
    I would like to see how the segments have been distributed.

  2. If I use a columnstore key in a DW fact table, what is the best strategy for combining with shard keys?

  3. If I have row segments in columnstore tables, would this be a way to do fragmentation without having to declare shard keys?

2

re: 1., see information_schema.columnar_segments; there is not general documentation for it yet, but the column names are self-explanatory for the most part.

re: 2., see SingleStoreDB Cloud · SingleStore Documentation

re: 3., if you don’t declare a shard key, we shard internally to give even distribution of the data across partitions (and thus leaf nodes). Within partitions, columnstore data will be broken into segments.