Suppose I have 12 segments in my system, then suppose I create a table where I distribute by a column based on month of the year. Now suppose I have a second table and I do likewise with that table, simply using a create statement with DISTRIBUTED BY (month)
and then I join these two tables on month as well as other attributes.
In the highly specific above scenario, is the performance of the query joining on month improved by the selection of month as a distribution column? Will records from October from both tables appear in the same segment auto-magically, or is it possible to specify the segment that these records will be distributed to such that they are collocated on the same segment? If they were and could be collocated, would there be any performance advantage anyway?