• fix an example after dependency change.
  • fix runner(..., na_pad) for vectors to return NA when windows is incomplete. Other methods already consistent.
  • fix the problems when calling runner::runner using do.call. (#83 and #84)
  • Fix runner.grouped_df (dplyr class) to not ignore k argument.
  • removed defunct argument type.
  • defunct type argument in favor of simplify.
  • fixed error when using runner::runner.
  • fixed max_run to return NA instead of NaN in incomplete or NA windows.
  • implement parallel windows computation.
  • added simplify identical to argument used in sapply.
  • add runner for xts objects.
  • fix runner for matrix.
  • type argument is being deprecated.
  • reduce number of dependencies.
  • add run_by function to prespecify arguments for multiple runner calls
  • runner supports grouped_df
  • runner function to be applied on data.frame or matrix - running windows constructed by subsetting rows.
  • runner output can be a list.
  • at lag and k allows to specify values as POSIXt sequence increment - same as by in seq.POSIXt.
  • enhanced function documentation by images.
  • runner function with default type = "auto" which automatically guess the type of returned object.
  • speed up runner function.
  • added at argument to all functions to return output with specific indexes.
  • switch to c++11 compiler
  • fix lag to get latest value within window instead of outside of the window
  • change argument in lag from k to lag
  • runner able to return output specified by user in type argument
  • allow negative lag
  • na_pad in all functions with lag or k
  • modified runner to accept any types of input vectors
  • added lag to all runner functions
  • added minmax_run which calculates running minimum and maximum to denote current highs and lows
  • improved vignettes
  • added runner function which allows to apply custom function on running windows - so far returning only numeric
  • all functions have additional idx argument which allows to compute running windows within specified date/time/indexes range.
  • added lag_run function - lagged value able to be computed on date/time lag.
  • added length_run function to compute number of elements within specified date/time/indexes range.