Change space in as.data.frame.table1()
(#60).
Add NMISS to output of stats.default()
for categorical variables.
Handle NMISS properly in
stats.apply.rounding() (no decimals).
Fix handling of abbreviated code with
render.missing.
Correct default rounding of PCTnoNA (ie, using
digits.pct).
Fix issue with t1kable (#82).
Fix a problem with t1kable when the row with the
label contains data.
Choose between qflextable, flextable
and regulartable in t1flex.
Add extra columns to groupspan headers in t1flex and t1kable (#93).
Option to place overall on the left (#88).
Missing values in stratification variable(s) raises and error (#80).
Add option rounding.fn to
stats.apply.rounding() to allow changing the rounding
behaviour (thanks to Niels Hagenbuch for the suggestion)
Add geometric standard deviation (GSD) to the output of
stats.default().
Fixed error when passing digits (thanks to Niels
Hagenbuch for reporting this).
Fixed integer formatting (issue #55).
Fixed issue #53.
Headline feature: ability to convert table1 output
to a data.frame, flextable or
kableExtra object (though with some limitations). As a
direct consequence, table1 can now be used within LaTeX
documents, and there is a new vignette to document this feature. Output
to .docx is also possible with this feature (through pandoc or
flextable).
Fomulas containing dot (‘.’) are now handled correctly (the ‘.’ means “all variables other than those that appear elsewhere in the formula”).
Arguments to formatC (e.g. big.mark) will apply to all numbers by default (previously, only continuous variables had this).
New setLabel() function provides an alternative
syntax for setting labels (useful for setting a label inside a
formula).
The headline feature is the extra.col option, which
gives a flexible way to add extra columns (e.g. P-values, SMD, …) to the
table (thanks to Matt Leipzig for helping out with suggestions and
testing on this).
Footnotes now appear in a
tag for better placement.
General improvements to the documentation.
Fixed issue that was causing vignette build to fail (due to changes made to the MatchIt package’s lalonde data set).
Allow multiple footnotes.
Added caption argument to generate a caption for the
table.
Set render.missing=NULL to ignore missing
values.
Added SUM to stats.default for
continuous variables.
Added PCTnoNA to stats.default and
na.is.category option to
render.categorical.default.
Changed (n=XX) to (N=XX) in default table header.
Improved error handling in function eqcut.
Function signif_pad can now use different decimal
marks to support different locales.
Added function round_pad (similar to
signif_pad but for fixed rounding).
Added function t1read.
Added quantile.type argument to
stats.default to allow selecting which quantile type is
computed.
Special handling for “0%” and “100%” in
stats.apply.rounding so that no decimals are
shown.
droplevels now works properly with nested
grouping.
table1 now returns a character string (containing
HTML). A print method handles the printing. There is no longer a need to
use capture.output to embed table1 in a shiny
app.
Removed standalone option as it is no longer
needed.
Use htmltools functionality to print.
Preview in RStudio notebook now works (issue #7).
Added knit_print method. No need to specify
results="asis" chunk option anymore when knitting.
Default CSS is included automatically. It can still be overridden by specifying a custom stylesheet in R Markdown.
Changed notation from upper case “Q” to small case “q” for
percentiles in stats.default.
Q1 and Q3 are now aliases for
q25 and q75 respectively in
stats.default.
Tertiles added to stats.default.
New function eqcut for creating equal-sized
categories from continuous variables.
New footnote argument for adding a footnote to the
table.
The vignette has been expanded with new examples and the documentation improved.