timestamp
as dimension produced unnecessarily warning due to checks; this has been fixed.stringsAsFactors
default configuration in data.frame
.The conversion to Date
or POSIXct
is not forced anymore & depends on the initial class. This brings significant speed to all functions (provided the initial class is already date/timestamp);
There have been some other minor improvements in speed, e.g. applying basic transformations to columns through data.table
’s set
instead of lapply
and .SD
.
The collapse_ranges
function got two new arguments, i.e. startAttr
and endAttr
- see ?collapse_ranges
;
Moreover, the core function has been re-written in Rcpp
for some speed gains.