Parallel Query: Query Plan: parallel table scan + group

DRAFT

We need to start making parts of the query plan parallel.

A good first step would be:

If we have a query plan in form:

  large table scan;

  temp-table-based grouping;

We can parallelize it like so:

  Distribute-to-workers(large table scan);

  Merge group-by-table(workers);

A parallel worker would get

• a table to scan (MDEV-39638, MDEV-39489)
• Scan chunk bounds (MDEV-39491)
• an Item* condition to apply
• .. todo something that copies data from source fields into temptable fields.
• a group-by temporary table to put result into
• Aggregate functions whose values to update.