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
i had removed these to make the dicts/sets more compact, but it would be nice to support efficiently calculating ranks/quantiles as well as iterating from them (e.g. iterate values between 25th and 75th quantile).
could this be set up to be decided at runtime? maybe e.g. by making two different nonempty types, one with length and one without, and specialize the NE constructor based on the types of left and right. updating length is just 1+length(left)+length(right)
The text was updated successfully, but these errors were encountered:
this turned out to be a very simple fix (just had to change two lines of code), but for a small key type (Int64), the size of the whole data structure will grow to 133% of previous size, for a feature that may or may not be useful. the main immediate use for length info would be to implement amount for #11 , but maybe that can be accomplished by storing length of the entire container (e.g. as an added field in RBDict/RBSet), which doesn't cause the same overhead.
i had removed these to make the dicts/sets more compact, but it would be nice to support efficiently calculating ranks/quantiles as well as iterating from them (e.g. iterate values between 25th and 75th quantile).
could this be set up to be decided at runtime? maybe e.g. by making two different nonempty types, one with length and one without, and specialize the
NE
constructor based on the types ofleft
andright
. updating length is just1+length(left)+length(right)
The text was updated successfully, but these errors were encountered: