Assume that needs to read the number of rows of about 100000,
Data distribution in several node and distribution in more than a dozen compared node, the performance will be how to change?


Or, to establish a partion key, is the only value choice is good, or select a category value is better?