You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
My apologies if this has already been brought up but do you plan on ever writing a rollapply function for data.table?
I know that you can use xts::rollapply function inside of data.table, however, xts::rollapply is still slow. I get the feeling that if a data.table::rollapply function was written utilizing data.table's radix sorting, it could be much faster than anything that is available right now. Without it, people just keep building c++ libraries to perform quick rolling time-series calcs...
The text was updated successfully, but these errors were encountered:
rollapply even when developed in data.table might not be as fast as custom built c++ code. Just because it is not possible to effectively optimize arbitrary R function.
closing as duplicate of #2778
My apologies if this has already been brought up but do you plan on ever writing a rollapply function for data.table?
I know that you can use xts::rollapply function inside of data.table, however, xts::rollapply is still slow. I get the feeling that if a data.table::rollapply function was written utilizing data.table's radix sorting, it could be much faster than anything that is available right now. Without it, people just keep building c++ libraries to perform quick rolling time-series calcs...
The text was updated successfully, but these errors were encountered: