High-Performance Open-Source Archive
\fontspec
instead of \global\setmainfont. The latter re-ran the math
font machinery on every cell, which could exhaust LaTeX math alphabets
and break compilation when a document contained many equations before
the table (#636).dim_pretty() height estimates now account for line
spacing and border widths.padding.left and padding.right are now
supported in PDF/LaTeX output.delete_rows() and delete_columns() no
longer reset all spans unconditionally. span_free() is now
only triggered when the deletion actually breaks a merged cell,
preserving existing merge structures in all other cases.as_image(),
colformat_image(), plot_chunk(),
gg_chunk(), grid_chunk()) gain an
alt parameter for alternative text. Alt text is rendered in
DOCX (descr attribute) and HTML (alt
attribute) output.fit_columns() constrains total table width
by shrinking columns proportionally. Text wraps inside narrower cells;
font sizes are unchanged. Columns that cannot shrink below their longest
word are clamped at that floor and remaining space is iteratively
redistributed among unclamped columns. A no_wrap argument
lets specific columns keep their optimal width.paginate() gains a new option
"starts" for argument group_def. When used,
group is an integer vector of body row indices where new
groups begin; page breaks are allowed before these rows.split_columns() and
split_rows() to split a flextable into a list of flextables
that fit within a given width or height. split_columns()
accepts column names, indices, or formulas for rep_cols,
allowing users to choose which columns to repeat and in what order.
split_rows() supports a group argument to keep
row groups together across pages; header and footer are repeated on
every page.split_to_pages() combines
split_rows() and split_columns() in a single
call for convenient two-dimensional pagination.as_flextable() for ‘rtables’
TableTree and ElementaryTable objects. The
conversion maps formatted content, column spans, alignments, indentation
and footnotes to flextable features. Use split_to_pages()
on the result for pagination.wrap_flextable() enables integration with
‘patchwork’ layouts. Flextable objects can be combined with ‘ggplot2’
plots using +, |, and /
operators. Table headers and footers are aligned with plot panel areas.
The panel argument controls alignment ("body",
"full", "rows", "cols") and the
space argument controls sizing ("free",
"fixed", "free_x", "free_y").
flex_body stretches body rows to match a neighbouring
plot’s panel height; flex_cols stretches data columns to
match the panel width. just controls horizontal alignment
("left", "right", "center") when
the table is narrower than the panel. S3 methods
ggplot_add.flextable and as_patch.flextable
are registered so that plot + flextable works
transparently.theme_borderless() applies a minimal theme
with no borders, bold header, and standard column alignment.fp_text_default().as_strike() to apply strikethrough
formatting to text chunks.compact_summary() to create a compact
summary of a data.frame that can be transformed as a flextable with
as_flextable().summarizor(): when using overall_label
with multiple by columns, an overall level is now added for
each grouping column (not only the last one). This produces margins at
every nesting level, including a grand total.footnote() gains a symbol_sep argument to
insert a separator between multiple footnote symbols in the same cell
(#699).as_qmd() to embed Quarto markdown
(cross-references, bold/italic, links, math, inline code) inside
flextable cells. Works with HTML, PDF and Word outputs. New function
use_flextable_qmd() installs the companion
flextable-qmd Lua filter extension in the Quarto
project.longtable environment, which only supports page breaks
between rows, not within a single row. HTML and Word outputs are not
affected.gdtools::strings_sizes() instead of
m_str_extents() and str_metrics(), goal is to
let ‘gdtools’ use only ‘systemfonts’ and be simplified.no-shadow-dom are no longer moved into a
Shadow DOM, restoring visibility when the document contains LaTeX
equations (pagedown#332).\n) in captions now render correctly in
PDF/LaTeX output (#663).word_style for a paragraph style works
now. The word_style values will be ignored if flextable is
process by ‘rmarkdown’ or ‘quarto’.as_flextable.tabulator(): the N= counts in column
headers are now displayed when there are multiple grouping columns
(previously limited to a single grouping column).footnote() no longer errors when the row selector
i matches zero rows (#712).fontspec LaTeX package is no longer
included when the PDF engine is pdflatex, fixing
compilation errors in Quarto documents using
pdf-engine: pdflatex (#701, #707). Engine detection now
also reads QUARTO_EXECUTE_INFO (Quarto >= 1.8) and
nested YAML (format > pdf > pdf-engine).none) in Quarto
output.merge_v(): vertically merged cell labels now appear at
the top of the merged range in PDF/LaTeX output instead of the bottom
(#654).valign) in merged cells now works
correctly in PDF/LaTeX output when rows have different heights (#639).
Content is placed in the first (top), middle (center), or last (bottom)
row of the merged range; \multirow is no longer used as it
miscalculates offsets with unequal row heights.by of summarizor() referring to two
columns, one of which has only one unique value no longer causes and
error when passed on to as_flextable().set_table_properties(opts_word = list(repeat_headers = FALSE)).format_fun.default so that it works with logical
columns.print.flextable(preview = "log") use str()
to show first values of data instead of print() so that
when there are ggplot2 v4 objects in the table, the print is not
failing.proc_freq gains new argument
count_format_fun to let control the function that format
the count values.0.9.8.as_raster(),
lollipop() and set_formatter_type().'' with
footnotes. Users should use add_footer_lines()
instead.tab_settings() is now using j argument as expected
(#635)set_table_properties() with
layout that defaults to “fixed”.delete_column() (#676)fmt_signif_after_zeros() fixed issue with 0
roundingproc_freq supports now non syntactically namesfix_border_issues function is now useless for
users, as it is now called automatically before printing.flextable_to_rmd() issue with images in pdf (issue
#651)flextable_to_rmd() issue with local chunk
eval option (issue #631)proc_freq can now display only the table percentages
without the count using include.table_count = FALSE.pagedown >= 0.20.2headers_flextable_at_bkm() and
footers_flextable_at_bkm() are defunct.flextable_to_rmd() is now using
knit_child() for safer usage from for loops or
if statements.save_as_image() and ph_with.flextable().as_raster() since gen_grob() is
easier to use and render nicer.align(), the default argument value
for align is now "left", rather than
c("left", "center", "right", "justify"). This returns the
default value to how it was in older versions of {flextable}.
align(), use of the old default align
argument could cause an error if the number of columns being adjusted
was not a multiple of 4.align had to be a
single value, when it could actually accept multiple values. This is why
a default value of c("left", "center", "right", "justify"),
was problematic. This documentation has now been updated and new
examples included in the documentation.align argument will now apply left
alignment to all columns in the body.path argument now has a signature of
part = c("body", "header", "footer", "all"), but
because only a single value can be selected, it will pick
"body" by default, as before.lollipop() since it produces (ugly) results
that can be replaced by nice results with gg_chunk() or
grid_chunk().as_image() when the table contains no
text.align() issue with recycling and update
documentation that was wrong about argument align that is
vectorized over columns.tab_settings() to set tabulation marks
configuration for Word and RTF. It works with
officer::fp_tabs().fmt_signif_after_zero() to round
significant figures after zeros.summarizor() don’t stop anymore if table only contain
discrete columns.as_flextable.data.frame() supports ‘data.table’footnote() handle undefined ref_symbols
argumentdelete_rows() does not delete rows if the row selection
is emptygen_grob() alignments when wrapping textgen_grob() when cells
are vertically mergedset_caption() can have no
bookmark and have autonumber used together.tabcolsep is now set to 2.set_formatter_type().fmt_2stats() so that it uses global flextable
settings, i.e. digits, etc.set_table_properties(opts_pdf = list(footer_repeat = TRUE)).as_chunk()ph_with.flextable() formats widths and heights
correctly.delete_columns() and
delete_rows() to let users delete rows or columns.save_as_image() now supports svg export with
‘svglite’.summarizor() function has been enhanced to offer
three new options:
by argument meaning ‘no grouping,’as_flextable.data.frame() always shows the number of
rows even if less than 10 (because I need it!).as_flextable.lm(),
as_flextable.gam(), as_flextable.glm(),
as_flextable.merMod() and as_flextable.htest()
now respect the global value of
getOption("show.signif.stars").add.random in
as_flextable.merMod() to let add or not random effects in
the final table.max_iter of function fit_to_width
is not ignored anymore.part to .part so
that column named part can be used.dim_pretty() returns correct numbers when not
‘inches’as_flextable.table() now propagates ... as
expectedset_table_properties(opts_pdf = list(caption_repeat = FALSE)).as_flextable() now works on tabular objects and
R < 4.1.to_html(type = "img") now use the correct width and
height1.4.as_flextable.tabular() now generate tabulated content
only if the sub group contains more than a single row. When possible,
row titles are tabulated.footers_flextable_at_bkm and
headers_flextable_at_bkm are deprecated. Instead, we
recommend using the prop_section() and
block_list() functions from the officer
package as alternative solutions. These functions provide more reliable
and flexible options for managing footers and headers.keepnext and ft.keepnext are
defunct, they are replaced by function paginate() that
enable nice pagination of tables across pages or
keep_with_next() to get total control over rows
pagination.set_flextable_defaults() gains parameter
pct_digits (number of digits for percentages) that will be
used in relevant functions (only proc_freq() for now).as_flextable.table().fmt_dbl(), fmt_int() and
fmt_pct().paginate()type is now possible when using
tabulator().save_as_html() ugly default title.set_header_labels() from a
simple un-named vector.set_formatter() now accepts single function to
be applied to all columns.gdtools::register_gfont().save_as_html(): use rmarkdown and add google
fonts if possible (See gdtools::register_gfont()).as_flextable.tabular(spread_first_col=TRUE)In short:
tables::tabular(): “Computes a table of
summary statistics, cross-classified by various variables”officer::rtf_add().tables::tabular() to flextable with new
function as_flextable.tabular().to_html.flextable() to make easy embedding of
flextable results in HTML (with ‘ggiraph’ for example).border.width (see
?set_flextable_defaults()) and set its default value to
.75, this setting is used in theme functions. The old default value was
hard coded to 1 and can be defined during the whole R session with
set_flextable_defaults(border.width = 1).officer_url_encode() to encode URL in office
files.knit_print.flextable() and related
functions.headers_flextable_at_bkm() and
footers_flextable_at_bkm()0.8.4.body_add_flextable().expand_single in
as_flextable.tabulator(). If FALSE, groups
with only one row will not be expanded with a title row.labels of labelizor() now support
functions in addition to named vectors.as.character() now returns the HTML string of the
tableas_flextable for data.frame using
df_printer() function implementation.expand_single in
as_grouped_data(). If FALSE, groups with only one row will
not be expanded with a title row.fmt_avg_dev(),
fmt_header_n(), fmt_n_percent() and renaming
of fmt_2stats() to fmt_summarizor() to help
working with tabulator(). tabulator() has also
new internal values that allow N=xxx notes and automatic labels.set_table_properties() is now the recommanded
way to set arguments related to format options and alignment in a
document. It is supposed to replace “knitr” chunk options
ft.align, ft.split, ft.keepnext,
ft.tabcolsep, ft.arraystretch,
ft.latex.float, ft.shadow,
fonts_ignore although they are all still supported. This
allows less interaction with the ‘R Markdown’ or ‘Quarto’ eco-system and
let to define it globally with
set_flextable_defaults().set_table_properties() and providing a scroll
value for argument opts_html:
opts_html = list(scroll = list(height = "500px",freeze_first_column = TRUE)).grid_chunk() to let users add grid
graphicsadd_header_row(),
add_footer_row(), add_body_row(),
add_header_lines() and add_footer_lines() now
supports formatted paragraph made with as_paragraph().as_paragraph()summarizor() can now be transformed directly as a
flextable with method as_flextable().body_add_flextable()prepend_chunks()line_spacing() as it is
expected to be a ratioft.keepnext now default to FALSE as lot of users had
issues with this option.xtable_to_flextable() is removed (use
as_flextable())gen_grob().labelizor() is a new function to help change text by
labels.fp_par())
in captions with set_caption().as_paragraph()set_table_properties() or
setting values to knitr chunk options tab.alt.title and
tab.alt.description.set_caption(). The alignment of the paragraph can be
different from the alignment of the table with the argument
align_with_table=FALSE.theme_apa (@rempsyc #426)as_flextable.tabulator() gains an argument
spread_first_col to enable spreading of the first column of
the table as a line separator.save_as_docx()
to avoid Word confusionfortify_width() calculationas_flextable.tabulator() gained an argument
label_rows used for labels to display in the first column
names, i.e. the row column names.shift_table() to produce Shift Tables used
used in clinical trial analysis ready to be used by
tabulator().as_image() don’t need anymore parameters
width and height if package ‘magick’ is
available.plot.flextable now default to grid Graphics.
It produce a plot object that can be used with packages ‘ggplot2’,
‘patchwork’ and ‘cowplot’. The raster version made with ‘webshot’ and
‘magick’ pachages is still available, use
plot(..., method = "webshot").add_footer_lines() instead.as_flextable() has now methods for lm, glm,
models from package ‘lme’ and ‘lme4’, htest (t.test, chisq.test, …),
gam, kmeans and pam.use_model_printer() to set model automatic
printing as a flextable in an R Markdown document.add_body_row() to add a row in the body
part with eventually merged/spanned columns.tabulator_colnames() to get column names
of a tabulator object.prepend_chunks() to prepend chunks of
content in flextable cells as with as_paragraph() but
without replacing the whole content.as_word_field() to let add
‘Word’ computed values into a flextable, as
as_word_field(x = "Page") for a page number.separate_header() to Separate collapsed
colnames into multiple rows.autofit() and dim_pretty() now
support newlines.autofit() and dim_pretty() now have an
argument hspans to help specify how horizontally spanned
cells should affect the results.ft.latex.float, supported values are ‘none’
(the default value), ‘float’, ‘wrap-r’, ‘wrap-l’, ‘wrap-i’ and
‘wrap-o’.ft.htmlscroll = FALSE.tabulator() can display any columns before and after
the columns of displayed aggregations.i, j and
part at the end of the function arguments, after the
....supp_data in the result table of
tabulator()j and source when
color is a function (in bg(), highlight() and
color()) is now based on colnames to avoid mistake while
mapping. #395append_chunks() to append chunks of
content in flextable cells as with as_paragraph() but
without replacing the whole content.tabulator() and its method
as_flextable.tabulator() that help the creation of tables
used in life science industry. It also comes with function
summarizor() and sugar function
fmt_2stats().empty_blanks() gained arguments width so
that users can also set blank columns’ width.... to format() function when using
colformat_num(). It makes possible to use arguments for
format(), for example
colformat_num(drop0trailing = TRUE).ft.keepnext and parameter
keepnext to function body_add_flextable() that
enable the Word option ‘keep rows together’, so that page break within a
table is avoided when possible.add_latex_dep() to manually add flextable
latex dependencies to the knitr session.<img> entries with invalid width attributes
e.g. width=40\%.df_printer that can be used via the
df_print option of R Markdown documents or by calling
use_df_printer().tab.cap.fp_text to
let format caption prefix in function
opts_current_table().colformat_* and set_flextable_defaults.surround() to ease the highlighting of
specific cells with borders.print(x, preview ="pdf").tab.topcaption that let modify the table
caption position from “top” (TRUE) to “bottom” (FALSE).before() to ease addition of
hline() before some values to match with.ft.shadow = TRUE to htmltools_value so
that shadow dom can not be used.fontname.set_flextable_defaults(line_spacing=1))as_equation for ‘MathJax’ equations.text_after to function
flextable_to_rmd to let append any text to the produced
flextable.chunk_dataframe for those who want to create
functions that work with as_paragraph.flextable-shadow-host.set_flextable_defaults now accept argument
padding that set values for padding top, bottom, left and
right.colorize, as_highlightnrow_part and ncol_keys are now
exportedcompose has a new argument use_dot to let
use . and loop over columnsinit_flextable_defaults()git diffcolformat_num calls the format
function on the numeric values (integer and float) which are therefore
displayed as in console R. This function is used during the creation of
a flextable so that by default the content of the cells is the same as
that displayed in console R.merge_v has a new argument combine to let
use j columns be used as a single value (all values are pasted).add_body for adding rows into a flextable
bodycolformat_image for images in
flextableas_flextable for gam
modelsset_flextable_defaults gained 4 new arguments
post_process_pdf, post_process_html,
post_process_docx and post_process_pptx to
enable flextable post-treatments conditionned by the output format.fp_text_default and
fp_border_default.\n and \t usage for pdfhighlight() for text highlighting
colorset_flextable_defaults() to set some
default formatting properties, i.e. default border color, font color,
padding, decimal.mark …save_as_docx gained a new argument
pr_section to define page layout with section properties,
save_as_html can now output more than a single table.colformat_ functions now use default values and filter
columns that are irrelevant (i.e. if colformat_num, only numeric values
are formatted). Also, new colformat_ functions have been
implemented (colformat_date,
colformat_datetime and colformat_double).plot_chunk and gg_chunk to
add miniplots or ggplots into a flextableph_with_flextable().fix_border_issues is the last instruction of all theme
functions so that borders are corrected if some cells have been
merged.tab_props$cap has a value.encodingline_spacing() (or with
function style()).i = NULL or
j = NULL, to select none, i = integer(0) or
j = integer(0).footnote is now able to add inline
footnotesxtable_to_flextable())continuous_summary(): summarize
continuous columns in a flextableautofit can now use only some parts of the
tables. This allows for example to no longer have gigantic columns by
not taking into account the “footer” part that is often composed of long
texts.set_table_properties lets adapt flextable
size as “100%”, “50%” of the available width for Word and HTML.bg() and color() now accept functions
(i.e. scales::col_numeric())display()as_chunk
(#152)officer::fp_sign importation that was not used
anymore so that officer can drop digest dependency.display().ph_with_flextable_at()knit_to_wml() has new arguments
align, split and
tab.cap.stylehtmltools_value() has a new argument
ft.alignflextable_html_dependency to get flextable
htmltools::htmlDependancy. This is necessary to output flextables in
html R Markdown documents from loop or other nested operations.save_as_docx,
save_as_pptx that lets users export flextable objects to
PowerPoint or Word documents.hrule to control how row heights should be
understood (at least, auto, exact)lollipop that lets users add mini lollipop
chart to flextable (kindly provided by github.com/pteridin)proc_freq got a new argument
weight to enable weighting of results.as_flextable.grouped_data() has now an
argument hide_grouplabel to let not print the group
names.webshot in
function save_as_image or with chunk option
webshot="webshot2".knit_to_wml to let display flextables from
non top level calls inside R Markdown document.ph_location* usage and make placement into slides
easier.fit_to_width to fit a flextable to a
maximum widthset_caption can now be used with R Markdown for Word
document and caption style can be defined with chunk option
tab.cap.style.save_as_image with R for Windowsplot),
as an image (see save_as_image) and raster for ggplot2 (see
as_raster).footnote to ease footnotes managementvalign to align vertically paragraphs in
cellproc_freq that mimic SAS proc freq
provided by Titouan Robert.linerange to produce mini lineranges.set_footer_dftheme_alafoli() and
theme_vader()align_text_col() and
align_nottext_col() to align columns by data typemerge_h_range() to merge a set of columns
row by rowfix_border_issues() fix issues with
borders when cells are mergedadd_header_row(),
add_footer_row(), add_header_lines() and
add_footer_lines() to add easily data in header or
footer.as_flextable() to let develop new
flextable functionsas_grouped_data() and its method
as_flextable() to create row titles to separate data in a
flextable.values for functions
set_header_labels and set_formattercompose that will replace
displayset_caption only available for html
outputknit_print() can be used with rmarkdown when rendering
to PowerPoint.regulartable and logical columnsbody_replace_flextable_at_bkm to
replace a bookmarked paragraph by a flextable.colformat_* to make content formatting
easier. It also deals with NA.format.flextable method so that users can
create their components.ft.align to align tables in
rmarkdown::word_document and ft.split to
activate Word option ‘Allow row to break across pages’.display()vline()\n is now supported. Function
autofit and dim_pretty do not support soft
returns and may return wrong results (will be considered as ““).border_outer()hyperlink_text() to be used with
display, font()hline*() and vline*() and
many new helper functions to be used instead of borders.xtable_to_flextable() that is returning a
flextable from an xtable object.htmltools_value() is exported for shiny
applications.knit_print() to render flextable in
rmarkdown.regulartable, faster and simpler than
flextableph_with_flextable_at to add a flextable at
any position in a slide.merge_at is providing a general way of
merging cells.theme_box()
Need mirroring services?
Contact our team at info@vpspulse.com.
Mirror powered by VPSpulse
Infrastructure sponsored by VPSPulse & Secure Payments by ArionPay.