Thanks for the nice extension! It is a good example of how to move development in this area.
Your code gave me an idea for reducing complexity in statistical extensions like pg_stat_statements. For testing purposes, I implemented an experimental base type, RunningStats, to uniformly calculate and store statistics in my pet project (https://github.com/danolivo/pg_track_optimizer).
But this project seems too simple to maintain such code. Maybe you see any sense in adding this type to your extension?
The general idea here is to have a column type that can store statistics on a parameter incrementally and compute the mean and stddev uniformly, as you do in your aggregate.
Thanks for the nice extension! It is a good example of how to move development in this area.
Your code gave me an idea for reducing complexity in statistical extensions like pg_stat_statements. For testing purposes, I implemented an experimental base type, RunningStats, to uniformly calculate and store statistics in my pet project (https://github.com/danolivo/pg_track_optimizer).
But this project seems too simple to maintain such code. Maybe you see any sense in adding this type to your extension?
The general idea here is to have a column type that can store statistics on a parameter incrementally and compute the mean and stddev uniformly, as you do in your aggregate.