Closed pentium3 closed 2 years ago
data pruning(剪枝): filtering out entries(from input data) that are guaranteed not to affect output (eg: duplicated entries). The database system then runs the same query but on the pruned data, which significantly reduces processing time.
Main idea: use the switch to prune a large portion of data based on the query, and the master only needs to process the remaining data in the same way as it does without the switch.
https://www.youtube.com/watch?v=ad9J0qpngrk