As per Vertica documentation, if the Group by Hash has to be avoided then the Projection should have the group by columns in the sort definition in the same order. But if the group by is on multiple tables how can we avoid group by Hash.
If you are trying to do grouping columns belonging to different tables after join, you may not be able to acheive group by pipeline unless your group by columns are columns used in join and you got mergejoin for join.
I am also noticing based on some questions you posted that you are trying to avoid hash join and group by hash . Is there any specific reason to do this ? In most cases hash join and group by hash works best.