library(officer)
# Package `magrittr` makes officer usage easier.
library(magrittr)

Quick start

  1. Start with read_docx

Use the function read_docx() to create an R object representing a Word document.

The initial Word file can be specified with the path argument. If none is provided, this file will be an empty document located in the package directory. Formats and styles are defined in the initial file.

From the initial document, we will be able to use an object containing not only paragraph styles, character styles and table styles of the original document but also its content.

my_doc <- read_docx() 
styles_info(my_doc)
##    style_type            style_id             style_name is_custom
## 1   paragraph              Normal                 Normal     FALSE
## 2   paragraph              Titre1              heading 1     FALSE
## 3   paragraph              Titre2              heading 2     FALSE
## 4   paragraph              Titre3              heading 3     FALSE
## 5   character      Policepardfaut Default Paragraph Font     FALSE
## 6       table       TableauNormal           Normal Table     FALSE
## 7   numbering         Aucuneliste                No List     FALSE
## 8   character              strong                 strong      TRUE
## 9   paragraph            centered               centered      TRUE
## 10      table       tabletemplate         table_template      TRUE
## 11      table Listeclaire-Accent2    Light List Accent 2     FALSE
## 12  character           Titre1Car            Titre 1 Car      TRUE
## 13  character           Titre2Car            Titre 2 Car      TRUE
## 14  character           Titre3Car            Titre 3 Car      TRUE
## 15  paragraph        graphictitle          graphic title      TRUE
## 16  paragraph          tabletitle            table title      TRUE
## 17      table       Professionnel     Table Professional     FALSE
## 18  paragraph                 TM1                  toc 1     FALSE
## 19  paragraph                 TM2                  toc 2     FALSE
## 20  paragraph       Textedebulles           Balloon Text     FALSE
## 21  character    TextedebullesCar    Texte de bulles Car      TRUE
##    is_default
## 1        TRUE
## 2       FALSE
## 3       FALSE
## 4       FALSE
## 5        TRUE
## 6        TRUE
## 7        TRUE
## 8       FALSE
## 9       FALSE
## 10      FALSE
## 11      FALSE
## 12      FALSE
## 13      FALSE
## 14      FALSE
## 15      FALSE
## 16      FALSE
## 17      FALSE
## 18      FALSE
## 19      FALSE
## 20      FALSE
## 21      FALSE
  1. Add elements to document

By default new content is added at the end of the document. To understand how to add content at any location in the document, see the later section about cursor manipulation.

Let’s create an image from a plot…

src <- tempfile(fileext = ".png")
png(filename = src, width = 5, height = 6, units = 'in', res = 300)
barplot(1:10, col = 1:10)
dev.off()

…and add that image to the document along with some new text paragraphs and a table.

my_doc <- my_doc %>% 
  body_add_img(src = src, width = 5, height = 6, style = "centered") %>% 
  body_add_par("Hello world!", style = "Normal") %>% 
  body_add_par("", style = "Normal") %>% # blank paragraph
  body_add_table(iris, style = "table_template")
  1. Write the Word file

An (updated) Word file can be generated using the print() function with the target argument:

print(my_doc, target = "assets/docx/first_example.docx")

Download file first_example.docx - view with office web viewer

Adding elements

There are two types of functions for adding elements.

  • add content as a paragraph: images, tables, text. Use body_add_* functions:
    • body_add_par
    • body_add_img
    • body_add_table
    • body_add_break
    • body_add_toc
    • body_add_gg
  • add text or image inside an existing paragraph. Use slip_in_* functions:
    • slip_in_img
    • slip_in_seqfield
    • slip_in_text

body_add_* functions

The paragraph is the main top container for content within a Word document. Note that tables are top container, they are at the same level as paragraphs. body_add_* functions are designed to add content as a top container: text as an entire paragraph, table, image, page break…

A title is a paragraph. To add a title, use body_add_par() with the style argument pointing to the corresponding title style.

Use the function styles_info() to see available styles:

read_docx() %>% styles_info() %>% 
  subset( style_type %in% "paragraph" )
##    style_type      style_id    style_name is_custom is_default
## 1   paragraph        Normal        Normal     FALSE       TRUE
## 2   paragraph        Titre1     heading 1     FALSE      FALSE
## 3   paragraph        Titre2     heading 2     FALSE      FALSE
## 4   paragraph        Titre3     heading 3     FALSE      FALSE
## 9   paragraph      centered      centered      TRUE      FALSE
## 15  paragraph  graphictitle graphic title      TRUE      FALSE
## 16  paragraph    tabletitle   table title      TRUE      FALSE
## 18  paragraph           TM1         toc 1     FALSE      FALSE
## 19  paragraph           TM2         toc 2     FALSE      FALSE
## 20  paragraph Textedebulles  Balloon Text     FALSE      FALSE

It is important to understand that these style names are read in the initial file provided to read_docx(). A few comments:

  • One of the styles is heading 1 which corresponds to a level 1 title.
  • When using body_add_gg() in the following code, using style = "centered" will set centered paragraph properties (defined as centered in the initial document) to the new paragraph where the plot will be added.
  • Table templates are also defined in the initial document and can be used with body_add_table(). For advanced tabular formatting, use the flextable package instead (flextable website). It has a function body_add_flextable() that can be used with officer.
if( require("ggplot2") ){
  gg <- ggplot(data = iris, aes(Sepal.Length, Petal.Length)) + 
    geom_point()
  
  read_docx() %>% 
    body_add_par(value = "Table of content", style = "heading 1") %>% 
    body_add_toc(level = 2) %>% 
    body_add_break() %>% 
  
    body_add_par(value = "dataset iris", style = "heading 2") %>% 
    body_add_table(value = head(iris), style = "table_template" ) %>% 
    
    body_add_par(value = "plot examples", style = "heading 1") %>% 
    body_add_gg(value = gg, style = "centered" ) %>% 
  
    print(target = "assets/docx/body_add_demo.docx") %>% 
    invisible()
}
## Le chargement a nécessité le package : ggplot2

Download file body_add_demo.docx - view with office web viewer

slip_in_* functions

The slip_in_* functions are designed to add content inside an existing paragraph: text, image or seq field. The element is inserted either at the beginning or end of the paragraph (pos = c('after', 'before')). Available functions are the following:

img.file <- file.path( R.home("doc"), "html", "logo.jpg" )
read_docx() %>%
  body_add_par("R logo: ", style = "Normal") %>%
  slip_in_img(src = img.file, style = "strong", 
              width = .3, height = .3, pos = "after") %>% 
  slip_in_text(" - This is ", style = "strong", pos = "before") %>% 
  slip_in_seqfield(str = "SEQ Figure \u005C* ARABIC",
    style = 'strong', pos = "before") %>% 
  print(target = "assets/docx/slip_in_demo.docx") %>% 
  invisible()

Download file slip_in_demo.docx - view with office web viewer

These have been implemented mostly to allow the addition of Word’s special sequence fields (which facilitate numbering) at the beginning of paragraphs used as reference entries (e.g. a table or plot caption). See the section Table and image captions.

Cursor manipulation

A cursor is available and can be manipulated so that content can be added relative to its position with the body_add_* functions:

  • before will insert a new element before the selected element in the document.
  • after will insert a new element after the selected element in the document.
  • on will replace the selected element in the document by a new element.

Cursor functions are the following:

In order to illustrate the cursor functions, a document made up of several paragraphs will be used (let’s use officer for that).

read_docx() %>%
  body_add_par("paragraph 1", style = "Normal") %>%
  body_add_par("paragraph 2", style = "Normal") %>%
  body_add_par("paragraph 3", style = "Normal") %>%
  body_add_par("paragraph 4", style = "Normal") %>%
  body_add_par("paragraph 5", style = "Normal") %>%
  body_add_par("paragraph 6", style = "Normal") %>%
  body_add_par("paragraph 7", style = "Normal") %>%
  print(target = "assets/docx/init_doc.docx" ) %>% 
  invisible()

Download file init_doc.docx - view with office web viewer

Now, let’s use init_doc.docx with read_docx() and manipulate its content with cursor functions.

doc <- read_docx(path = "assets/docx/init_doc.docx") %>%

  # default template contains only an empty paragraph
  # Using cursor_begin and body_remove, we can delete it
  cursor_begin() %>% body_remove() %>%

  # Let add text at the beginning of the
  # paragraph containing text "paragraph 4"
  cursor_reach(keyword = "paragraph 4") %>%
  slip_in_text("This is ", pos = "before", style = "Default Paragraph Font") %>%

  # move the cursor forward and end a section
  cursor_forward() %>%
  body_add_par("The section stop here", style = "Normal") %>%
  body_end_section(landscape = TRUE, continuous = FALSE) %>%

  # move the cursor at the end of the document
  cursor_end() %>%
  body_add_par("The document ends now", style = "Normal")

print(doc, target = "assets/docx/cursor.docx") %>% 
  invisible()

Download file cursor.docx - view with office web viewer

Remove content

The function body_remove() lets you remove content from a Word document. This function used with cursor_* functions is a convenient tool to update an existing document.

For illustration purposes, we will generate a document that will be used as an initial document later when showing how to use body_remove().

library(officer)
library(magrittr)

str1 <- "Lorem ipsum dolor sit amet, consectetur adipiscing elit. " %>% 
  rep(20) %>% paste(collapse = "")
str2 <- "Drop that text" 
str3 <- "Aenean venenatis varius elit et fermentum vivamus vehicula. " %>% 
  rep(20) %>% paste(collapse = "")

my_doc <- read_docx()  %>% 
  body_add_par(value = str1, style = "Normal") %>% 
  body_add_par(value = str2, style = "centered") %>% 
  body_add_par(value = str3, style = "Normal") 

print(my_doc, target = "assets/docx/ipsum_doc.docx") %>% invisible()

The file ipsum_doc.docx now exists and contains a paragraph containing text that text. In the following example, we will position the cursor on that paragraph and then delete it:

my_doc <- read_docx(path = "assets/docx/ipsum_doc.docx")  %>% 
  cursor_reach(keyword = "that text") %>% 
  body_remove()

print(my_doc, target = "assets/docx/ipsum_doc.docx") %>% invisible()

The text search is made via xpath 1.0 and regular expressions are not supported.

Download file ipsum_doc.docx - view with office web viewer

Replace content

The body_add_* functions let you replace content in a Word document.

For illustration purposes, we will generate a document that will be used as an initial document later.

my_doc <- read_docx()  %>% 
  body_add_par(value = str1, style = "Normal") %>% 
  body_add_par(value = str2, style = "centered") %>% 
  body_add_par(value = str3, style = "Normal") 

print(my_doc, target = "assets/docx/replace_template.docx")

The file replace_template.docx now exists and contains a paragraph containing text that text. In the following example, we will position the cursor on that paragraph and then replace it. Using pos = "on" will replace content where the cursor is with new content.

my_doc <- read_docx(path = "assets/docx/replace_template.docx")  %>% 
  cursor_reach(keyword = "that text") %>% 
  body_add_par(value = "This is a new paragraph.", style = "centered", pos = "on")

print(my_doc, target = "assets/docx/replace_doc.docx")

Search-and-replace

You can also use the body_replace_* functions to search-and-replace text. body_replace_at() replaces text at a bookmark:

doc <- read_docx() %>%
  body_add_par("centered text", style = "centered") %>%
  slip_in_text(". How are you", style = "strong") %>%
  body_bookmark("text_to_replace") %>%
  body_replace_at("text_to_replace", "not left aligned")

body_replace_all_text() will, depending on the options provided, replace text either at the cursor or in the entire document:

doc <- read_docx() %>%
  body_add_par("Placeholder one") %>%
  body_add_par("Placeholder two")

# Show text chunk at cursor
docx_show_chunk(doc)  # Output is 'Placeholder two'
## 1 text nodes found at this cursor.
##   <w:t>: 'Placeholder two'
# Simple search-and-replace at current cursor, with regex turned off
body_replace_all_text(doc, "Placeholder", "new", only_at_cursor = TRUE, fixed=TRUE)
docx_show_chunk(doc)  # Output is 'new two'
## 1 text nodes found at this cursor.
##   <w:t>: 'new two'
# Do the same, but in the entire document and ignoring case
body_replace_all_text(doc, "placeholder", "new", only_at_cursor = FALSE, ignore.case=TRUE)
cursor_backward(doc)
docx_show_chunk(doc) # Output is 'new one'
## 1 text nodes found at this cursor.
##   <w:t>: 'new one'
# Use regex : replace all words starting with "n" with the word "example"
body_replace_all_text(doc, "\\bn.*?\\b", "example")
docx_show_chunk(doc) # Output is 'example one'
## 1 text nodes found at this cursor.
##   <w:t>: 'example one'

Download file replace_doc.docx - view with office web viewer

Sections

Sections can be added to a document. This is possible by using the function body_end_section(). The default section can be modified with the function body_default_section().

A section starts at the end of the previous section (or the beginning of the document if no preceding section exists). It stops where the section is declared.

str1 <- "Lorem ipsum dolor sit amet, consectetur adipiscing elit. " %>% 
  rep(30) %>% paste(collapse = "")
str2 <- "Aenean venenatis varius elit et fermentum vivamus vehicula. " %>% 
  rep(30) %>% paste(collapse = "")

my_doc <- read_docx()  %>% 
  slip_in_text(str = str1, style = "strong") %>% 
  body_add_par(value = str2, style = "centered") %>% 
  break_column_before() %>% 
  body_end_section(continuous = TRUE, 
                   colwidths = c(.6, .4), space = .05, sep = FALSE) %>%
  body_add_par(value = str3, style = "Normal") 
print(my_doc, target = "assets/docx/section.docx") %>% invisible()

Download file section.docx - view with office web viewer

In the previous example, the first two paragraphs will be in a 2 column section and the third will be in a default section.

Table and image captions

slip_in_seqfield() and slip_in_text() can be combined to prefix a paragraph with references (i.e. chapter number and graphic index in the document). However, producing a plot or a table and its caption can be verbose.

Shortcut functions are implemented in the object shortcuts (it will at least give you a template of code to modify if it does not fit your needs exactly). slip_in_tableref(), slip_in_plotref() and body_add_gg() can make life easier.

Usage of these functions is illustrated below:

library(magrittr)
library(officer)
if( require("ggplot2") ){

gg1 <- ggplot(data = iris, aes(Sepal.Length, Petal.Length)) + 
  geom_point()
gg2 <- ggplot(data = iris, aes(Sepal.Length, Petal.Length, color = Species)) + 
  geom_point()


doc <- read_docx() %>% 
  body_add_par(value = "Table of content", style = "heading 1") %>% 
  body_add_toc(level = 2) %>% 
  
  body_add_par(value = "Tables", style = "heading 1") %>% 
  body_add_par(value = "dataset mtcars", style = "heading 2") %>% 
  body_add_table(value = head(mtcars)[, 1:4], style = "table_template" ) %>% 
  body_add_par(value = "data mtcars", style = "table title") %>% 
  shortcuts$slip_in_tableref(depth = 2) %>%
  
  body_add_par(value = "dataset iris", style = "heading 2") %>% 
  body_add_table(value = head(iris), style = "table_template" ) %>% 
  body_add_par(value = "data iris", style = "table title") %>% 
  shortcuts$slip_in_tableref(depth = 2) %>%
  
  body_end_section(continuous = FALSE, landscape = FALSE ) %>% 
  
  body_add_par(value = "plot examples", style = "heading 1") %>% 
  body_add_gg(value = gg1, style = "centered" ) %>% 
  body_add_par(value = "graph example 1", style = "graphic title") %>% 
  shortcuts$slip_in_plotref(depth = 1) %>%
  
  body_add_par(value = "plot 2", style = "heading 2") %>% 
  body_add_gg(value = gg2, style = "centered" ) %>% 
  body_add_par(value = "graph example 2", style = "graphic title") %>% 
  shortcuts$slip_in_plotref(depth = 2) %>%
  
  body_end_section(continuous = FALSE, landscape = TRUE) %>% 
  
  body_add_par(value = "Table of tables", style = "heading 2") %>% 
  body_add_toc(style = "table title") %>% 
  body_add_par(value = "Table of graphics", style = "heading 2") %>% 
  body_add_toc(style = "graphic title")

print(doc, target = "assets/docx/toc_and_captions.docx") %>% invisible()
}

Download file toc_and_captions.docx - view with office web viewer