- New functions
frequency_grid_df()
andfrequency_grid_plot()
to analyze and visualize possible frequency distributions given missing values. - Fixed a bug in
mode_count_range()
that displayed a maximum of 0 for certain kinds of input. - In
mode_possible_min()
andmode_possible_max()
, themultiple
argument was renamed toaccept
for greater consistency withmode_first()
and other functions. mode_possible_max()
no longer returnsNA
if the number ofNA
s in the input is equal to or higher than the number of "empty slots" (i.e., the hypothetical non-NA
values that would have to be present inx
in order for all of its unique values to be equally frequent). After all, it is possible for all unique values to be "filled up" byNA
s so that they are themselves modes.
Patch for CRAN submission.
mode_count()
and all other metadata functions now have amax_unique
argument, allowing users to encode knowledge about missing values in the analysis.mode_single()
now has anaccept
argument for opting into a less strict check for a single mode, in analogy tomode_first()
.mode_first()
had itsfirst_known
argument renamed toaccept
for consistency withmode_single()
.- Further improvements to the documentation, including vignettes.
- Fixed incongruencies in the documentation and in some warnings.
- Established
R CMD check
with Github Actions.
- New functions for modal metadata:
mode_is_trivial()
mode_frequency()
mode_count_range()
mode_frequency_range()
- New vignettes on missing values, metadata, and performance.
- Many bugfixes and other improvements.
- Added a
NEWS.md
file to track changes to the package.