Skip to contents

Function used to render flextable in knitr/rmarkdown documents.

You should not call this method directly. This function is used by the knitr package to automatically display a flextable in an "R Markdown" document from a chunk. However, it is recommended to read its documentation in order to get familiar with the different options available.

R Markdown outputs can be :

  • HTML

  • 'Microsoft Word'

  • 'Microsoft PowerPoint'

  • PDF

Table captioning is a flextable feature compatible with R Markdown documents. The feature is available for HTML, PDF and Word documents. Compatibility with the "bookdown" package is also ensured, including the ability to produce captions so that they can be used in cross-referencing.

For Word, it's recommanded to work with package 'officedown' that supports all features of flextable.


# S3 method for flextable
knit_print(x, ...)



a flextable object




Supported formats require some minimum pandoc versions:

Output formatpandoc minimal version
HTML>= 1.12
Word (docx)>= 2.0
PowerPoint (pptx)>= 2.4
PDF>= 1.12

Chunk options

Some features, often specific to an output format, are available to help you configure some global settings relatve to the table output. knitr's chunk options are to be used to change the default settings:

  • HTML, PDF and Word:

    • ft.align: flextable alignment, supported values are 'left', 'center' and 'right'. Its default value is 'center'.

  • HTML only:

    • ft.htmlscroll, can be TRUE or FALSE (default) to enable horizontal scrolling. Use set_table_properties() for more options about scrolling.

  • Word only:

    • ft.split Word option 'Allow row to break across pages' can be activated when TRUE (default value).

    • ft.keepnext defunct in favor of paginate()

  • PDF only:

    • ft.tabcolsep space between the text and the left/right border of its containing cell, the default value is 0 points.

    • ft.arraystretch height of each row relative to its default height, the default value is 1.5.

    • ft.latex.float type of floating placement in the document, one of:

      • 'none' (the default value), table is placed after the preceding paragraph.

      • 'float', table can float to a place in the text where it fits best

      • 'wrap-r', wrap text around the table positioned to the right side of the text

      • 'wrap-l', wrap text around the table positioned to the left side of the text

      • 'wrap-i', wrap text around the table positioned inside edge-near the binding

      • 'wrap-o', wrap text around the table positioned outside edge-far from the binding

  • PowerPoint only:

    • ft.left, Position should be defined with these options. Theses are the top left coordinates in inches of the placeholder that will contain the table. Their default values are 1 and 2 inches.

If some values are to be used all the time in the same document, it is recommended to set these values in a 'knitr r chunk' by using function knitr::opts_chunk$set(ft.split=FALSE, ...).

Table caption

Captions can be defined in two ways.

The first is with the set_caption() function. If it is used, the other method will be ignored. The second method is by using knitr chunk option tab.cap.

set_caption(x, caption = "my caption")

If set_caption function is not used, caption identifier will be read from knitr's chunk option Note that in a bookdown and when not using officedown::rdocx_document(), the usual numbering feature of bookdown is used.'my_id'.

Some options are available to customise captions for any output:

Word stylename to use for table
caption id/bookmarktab.idNULL
display table caption on top of the table or nottab.topcaptionTRUE
caption table sequence"tab:"

Word output when officedown::rdocx_document() is used is coming with more options such as ability to choose the prefix for numbering chunk for example. The table below expose these options:

prefix for numbering chunk (default to "Table ").tab.cap.preTable
suffix for numbering chunk (default to ": ").tab.cap.sep" :"
title number depthtab.cap.tnd0
caption prefix formatting propertiestab.cap.fp_textfp_text_lite(bold = TRUE)
separator to use between title number and table"-"

HTML output

HTML output is using shadow dom to encapsule the table into an isolated part of the page so that no clash happens with styles.

PDF output

Some features are not implemented in PDF due to technical infeasibility. These are the padding, line_spacing and height properties. Note also justified text is not supported and is transformed to left.

It is recommended to set theses values in a 'knitr r chunk' so that they are permanent all along the document: knitr::opts_chunk$set(ft.tabcolsep=0, ft.latex.float = "none").

See add_latex_dep() if caching flextable results in 'R Markdown' documents.

PowerPoint output

Auto-adjust Layout is not available for PowerPoint, PowerPoint only support fixed layout. It's then often necessary to call function autofit() so that the columns' widths are adjusted if user does not provide the withs.

Images cannot be integrated into tables with the PowerPoint format.


if (FALSE) {
if (pandoc_available() &&
  pandoc_version() > numeric_version("2")) {
  demo_loop <- system.file(
    package = "flextable",
  rmd_file <- tempfile(fileext = ".Rmd")
  file.copy(demo_loop, to = rmd_file, overwrite = TRUE)
    input = rmd_file, output_format = "html_document",
    output_file = "demo.html"