Chapter 8 Extract content

8.1 Import Word document

The function docx_summary() returns the content of a Word document.

library(officer)
example_docx <- system.file(package = "officer", "doc_examples/example.docx")
doc <- read_docx(example_docx)
content <- docx_summary(doc)
content

doc_index	para_id	content_type	style_name	text	table_index	row_id	cell_id	is_header	row_span	col_span	table_stylename
integer	character	character	character	character	integer	integer	integer	logical	integer	character	character
1	4248496C	paragraph	heading 1	Title 1
2	6A358887	paragraph		Lorem ipsum dolor sit amet, consectetur adipiscing elit.
3	22C38A25	paragraph	heading 1	Title 2
4	0358E421	paragraph	List Paragraph	Quisque tristique
5	70C2F11E	paragraph	List Paragraph	Augue nisi, et convallis
6	429784E2	paragraph	List Paragraph	Sapien mollis nec.
7	49D1AAF7	paragraph	heading 2	Sub title 1
8	2F68EB53	paragraph	List Paragraph	Quisque tristique
9	6718A3AB	paragraph	List Paragraph	Augue nisi, et convallis
10	374BD321	paragraph	List Paragraph	Sapien mollis nec.
12	0B98AFE8	paragraph		Phasellus nec nunc vitae nulla interdum volutpat eu ac massa.
13	7BA11984	paragraph	heading 2	Sub title 2
14	6E5CCCAB	paragraph		Morbi rhoncus sapien sit amet leo eleifend, vel fermentum nisi mattis.
16	105C9948	table cell		Petals	1	1	1	true	1	1	Light Shading
17	4B712A9C	table cell		Internode	1	1	2	true	1	1	Light Shading
18	4F7C7923	table cell		Sepal	1	1	3	true	1	1	Light Shading
19	75402514	table cell		Bract	1	1	4	true	1	1	Light Shading
20	0AA8CF0C	table cell		5,621498349	1	2	1	false	1	2	Light Shading
21	07F46688	table cell		2,462106579	1	2	3	false	1	2	Light Shading
22	058EDA18	table cell		18,2034091	1	2	4	false	1	2	Light Shading
23	529A35E8	table cell		4,994616997	1	3	1	false	1	1	Light Shading
24	0112F740	table cell		AA	1	3	2	false	2	1	Light Shading
25	4FA4104A	table cell		2,429320759	1	3	3	false	1	1	Light Shading
26	6411D002	table cell		17,65204912	1	3	4	false	1	1	Light Shading
27	4BE8A109	table cell		4,767504884	1	4	1	false	1	1	Light Shading
29	03C4A309	table cell		AAA	1	4	3	false	1	2	Light Shading
30	6742B786	table cell		25,9242382	1	5	1	false	1	2	Light Shading
31	058788C5	table cell		2,066051345	1	5	3	false	1	1	Light Shading
32	3B968F0F	table cell		18,37915478	1	5	4	false	1	1	Light Shading
33	502FD7C0	table cell		6,489375001	1	6	1	false	1	1	Light Shading
34	5674C826	table cell		25,21130805	1	6	2	false	1	1	Light Shading
35	592F4D1E	table cell		2,901582763	1	6	3	false	1	1	Light Shading
36	5A8BE950	table cell		17,31304737	1	6	4	false	1	1	Light Shading
37	448288CA	table cell		17,07215724	1	6	4	false	1	1	Light Shading
38	06AE66A5	table cell		18,2902189	1	6	4	false	3	1	Light Shading
39	22CBAAC4	table cell		5,7858682	1	7	1	false	1	1	Light Shading
40	7A62E4DC	table cell		25,52433147	1	7	2	false	1	1	Light Shading
41	3EF2CCF3	table cell		2,655642742	1	7	3	false	1	1	Light Shading
43	7F987E3C	table cell		5,645575295	1	8	1	false	1	1	Light Shading
44	182FB64A	table cell		Merged cell	1	8	2	false	4	1	Light Shading
45	7970FD56	table cell		2,278691288	1	8	3	false	1	1	Light Shading
47	5A3B1DFD	table cell		4,828953215	1	9	1	false	1	1	Light Shading
49	24C6DA47	table cell		2,238467716	1	9	3	false	1	1	Light Shading
50	7BB0A36A	table cell		19,87376227	1	9	4	false	1	1	Light Shading
51	6241A0F7	table cell		6,783500773	1	10	1	false	1	1	Light Shading
53	5D5B28B3	table cell		2,202762147	1	10	3	false	1	1	Light Shading
54	0D353C70	table cell		19,85326662	1	10	4	false	1	1	Light Shading
55	460CBC23	table cell		5,395076839	1	11	1	false	1	1	Light Shading
57	493EC235	table cell		2,538375992	1	11	3	false	1	1	Light Shading
58	07AD5415	table cell		19,56545356	1	11	4	false	1	1	Light Shading
59	74CD3ED9	table cell		4,683617783	1	12	1	false	1	1	Light Shading
60	00F0FDA0	table cell		29,2459239	1	12	2	false	1	1	Light Shading
61	383EA117	table cell		2,601945544	1	12	3	false	1	1	Light Shading
62	55B6D183	table cell		18,95335451	1	12	4	false	1	1	Light Shading
63	4E9DE916	table cell		Note	1	13	1	false	1	4	Light Shading
64	3FC8AA5E	table cell		New line note	1	13	4	false	1	4	Light Shading
n: 56

Explore the results:

count(content, content_type)

content_type	n
character	integer
paragraph	13
table cell	43
n: 2

To get all paragraphs:

filter(content, content_type %in% "paragraph") |>
  select(doc_index, style_name, text)

doc_index	style_name	text
integer	character	character
1	heading 1	Title 1
2		Lorem ipsum dolor sit amet, consectetur adipiscing elit.
3	heading 1	Title 2
4	List Paragraph	Quisque tristique
5	List Paragraph	Augue nisi, et convallis
6	List Paragraph	Sapien mollis nec.
7	heading 2	Sub title 1
8	List Paragraph	Quisque tristique
9	List Paragraph	Augue nisi, et convallis
10	List Paragraph	Sapien mollis nec.
12		Phasellus nec nunc vitae nulla interdum volutpat eu ac massa.
13	heading 2	Sub title 2
14		Morbi rhoncus sapien sit amet leo eleifend, vel fermentum nisi mattis.
n: 13

8.1.1 Run-level extraction

By default, docx_summary() aggregates content at the paragraph (or table cell) level: one row per paragraph, one column text with the concatenated content. Passing detailed = TRUE switches the function to a run-level data.frame: one row per run, with additional columns exposing text formatting, hyperlinks, bookmarks, footnotes, field codes and embedded image paths.

Both dplyr and tidyr must be available when detailed = TRUE.

To illustrate features that are not present in the bundled example.docx, the chunk below prepares a companion document with a bookmark, an external hyperlink, an internal link pointing to that bookmark, and a footnote.

doc_detailed <- read_docx(path = "static/office/example_detailed.docx")
runs <- docx_summary(doc_detailed, detailed = TRUE)
runs

doc_index	para_id	content_type	run_index	run_content_index	run_content_text	image_path	field_code	footnote_text	link	link_to_bookmark	bookmark_start	character_stylename	sz	sz_cs	font_family_ascii	font_family_eastasia	font_family_hansi	font_family_cs	bold	italic	underline	color	shading	shading_color	shading_fill	paragraph_stylename	keep_with_next	align	level	num_id	table_index	row_id	cell_id	col_span	row_span	is_header	table_stylename
integer	character	character	integer	integer	character	character	character	character	character	character	character	character	integer	integer	character	character	character	character	logical	logical	logical	character	character	character	character	character	logical	character	integer	integer	integer	integer	integer	character	integer	logical	character
1		paragraph	1	1	See														false	false	false					Normal	false
1		paragraph	2	1	the officer website				https://ardata-fr.github.io/officeverse/										false	false	false					Normal	false
1		paragraph	3	1	or jump to the														false	false	false					Normal	false
1		paragraph	4	1	anchor below				anchor										false	false	false					Normal	false
1		paragraph	5	1	for more details.														false	false	false					Normal	false
2		paragraph	1	1	A paragraph with a														false	false	false					Normal	false
2		paragraph	2	1				This is a footnote.					20	20	Arial	Arial	Arial	Arial	false	false	false	#000000				Normal	false
2		paragraph	3	1	footnote reference.														false	false	false					Normal	false
3		paragraph	1	1							anchor								false	false	false					Normal	false
4		paragraph	1	1	Target of the internal link.														false	false	false					Normal	false
n: 10

External hyperlinks surface in the link column, internal links in link_to_bookmark:

runs |>
  filter(!is.na(link) | !is.na(link_to_bookmark)) |>
  select(doc_index, run_content_text, link, link_to_bookmark)

doc_index	run_content_text	link	link_to_bookmark
integer	character	character	character
1	the officer website	https://ardata-fr.github.io/officeverse/
1	anchor below	anchor
n: 2

Bookmarks starting on a paragraph are concatenated (with |) in bookmark_start:

runs |>
  filter(!is.na(bookmark_start)) |>
  select(doc_index, bookmark_start, run_content_text)

doc_index integer	3
bookmark_start character	anchor
run_content_text character

Footnote contents are carried in footnote_text:

runs |>
  filter(!is.na(footnote_text)) |>
  select(doc_index, run_content_text, footnote_text)

doc_index integer	2
run_content_text character
footnote_text character	This is a footnote.

Run-level formatting (bold, italic, underline, color, character style, font size in half-points, font families, paragraph alignment, numbering level) is exposed in dedicated columns. For example, on the bundled example.docx:

runs_ex <- docx_summary(doc, detailed = TRUE)
runs_ex |>
  filter(bold %in% TRUE | italic %in% TRUE) |>
  select(doc_index, run_content_text, bold, italic, color, sz) 

doc_index	run_content_text	bold	italic	color	sz
integer	character	logical	logical	character	integer
n: 0

Embedded images are extracted to the temporary directory of the rdocx object and their path is available in the image_path column. If the run refers to an image, the other text-oriented columns are NA. Copy the files to a permanent location before the R session ends if you need them afterwards.

8.1.2 Word tables

Tables are unstacked:

table_cells <- filter(content, content_type %in% "table cell")
table_cells

doc_index	para_id	content_type	style_name	text	table_index	row_id	cell_id	is_header	row_span	col_span	table_stylename
integer	character	character	character	character	integer	integer	integer	logical	integer	character	character
16	105C9948	table cell		Petals	1	1	1	true	1	1	Light Shading
17	4B712A9C	table cell		Internode	1	1	2	true	1	1	Light Shading
18	4F7C7923	table cell		Sepal	1	1	3	true	1	1	Light Shading
19	75402514	table cell		Bract	1	1	4	true	1	1	Light Shading
20	0AA8CF0C	table cell		5,621498349	1	2	1	false	1	2	Light Shading
21	07F46688	table cell		2,462106579	1	2	3	false	1	2	Light Shading
22	058EDA18	table cell		18,2034091	1	2	4	false	1	2	Light Shading
23	529A35E8	table cell		4,994616997	1	3	1	false	1	1	Light Shading
24	0112F740	table cell		AA	1	3	2	false	2	1	Light Shading
25	4FA4104A	table cell		2,429320759	1	3	3	false	1	1	Light Shading
26	6411D002	table cell		17,65204912	1	3	4	false	1	1	Light Shading
27	4BE8A109	table cell		4,767504884	1	4	1	false	1	1	Light Shading
29	03C4A309	table cell		AAA	1	4	3	false	1	2	Light Shading
30	6742B786	table cell		25,9242382	1	5	1	false	1	2	Light Shading
31	058788C5	table cell		2,066051345	1	5	3	false	1	1	Light Shading
32	3B968F0F	table cell		18,37915478	1	5	4	false	1	1	Light Shading
33	502FD7C0	table cell		6,489375001	1	6	1	false	1	1	Light Shading
34	5674C826	table cell		25,21130805	1	6	2	false	1	1	Light Shading
35	592F4D1E	table cell		2,901582763	1	6	3	false	1	1	Light Shading
36	5A8BE950	table cell		17,31304737	1	6	4	false	1	1	Light Shading
37	448288CA	table cell		17,07215724	1	6	4	false	1	1	Light Shading
38	06AE66A5	table cell		18,2902189	1	6	4	false	3	1	Light Shading
39	22CBAAC4	table cell		5,7858682	1	7	1	false	1	1	Light Shading
40	7A62E4DC	table cell		25,52433147	1	7	2	false	1	1	Light Shading
41	3EF2CCF3	table cell		2,655642742	1	7	3	false	1	1	Light Shading
43	7F987E3C	table cell		5,645575295	1	8	1	false	1	1	Light Shading
44	182FB64A	table cell		Merged cell	1	8	2	false	4	1	Light Shading
45	7970FD56	table cell		2,278691288	1	8	3	false	1	1	Light Shading
47	5A3B1DFD	table cell		4,828953215	1	9	1	false	1	1	Light Shading
49	24C6DA47	table cell		2,238467716	1	9	3	false	1	1	Light Shading
50	7BB0A36A	table cell		19,87376227	1	9	4	false	1	1	Light Shading
51	6241A0F7	table cell		6,783500773	1	10	1	false	1	1	Light Shading
53	5D5B28B3	table cell		2,202762147	1	10	3	false	1	1	Light Shading
54	0D353C70	table cell		19,85326662	1	10	4	false	1	1	Light Shading
55	460CBC23	table cell		5,395076839	1	11	1	false	1	1	Light Shading
57	493EC235	table cell		2,538375992	1	11	3	false	1	1	Light Shading
58	07AD5415	table cell		19,56545356	1	11	4	false	1	1	Light Shading
59	74CD3ED9	table cell		4,683617783	1	12	1	false	1	1	Light Shading
60	00F0FDA0	table cell		29,2459239	1	12	2	false	1	1	Light Shading
61	383EA117	table cell		2,601945544	1	12	3	false	1	1	Light Shading
62	55B6D183	table cell		18,95335451	1	12	4	false	1	1	Light Shading
63	4E9DE916	table cell		Note	1	13	1	false	1	4	Light Shading
64	3FC8AA5E	table cell		New line note	1	13	4	false	1	4	Light Shading
n: 43

Cells positions and values are dispatched in columns row_id, cell_id, text and is_header (a logical column indicating if the cell is part of a header or not). Note that the content itself (column text) is a character vector.

table_body <- table_cells |>
  filter(!is_header) |>
  select(row_id, cell_id, text)
table_body

row_id	cell_id	text
integer	integer	character
2	1	5,621498349
2	3	2,462106579
2	4	18,2034091
3	1	4,994616997
3	2	AA
3	3	2,429320759
3	4	17,65204912
4	1	4,767504884
4	3	AAA
5	1	25,9242382
5	3	2,066051345
5	4	18,37915478
6	1	6,489375001
6	2	25,21130805
6	3	2,901582763
6	4	17,31304737
6	4	17,07215724
6	4	18,2902189
7	1	5,7858682
7	2	25,52433147
7	3	2,655642742
8	1	5,645575295
8	2	Merged cell
8	3	2,278691288
9	1	4,828953215
9	3	2,238467716
9	4	19,87376227
10	1	6,783500773
10	3	2,202762147
10	4	19,85326662
11	1	5,395076839
11	3	2,538375992
11	4	19,56545356
12	1	4,683617783
12	2	29,2459239
12	3	2,601945544
12	4	18,95335451
13	1	Note
13	4	New line note
n: 39

Reshaping the data with columns row_id, cell_id and text would display something close to the orginal table:

table_body |>
  pivot_wider(names_from = cell_id, values_from = text)

# Warning: Values from `text` are not uniquely identified; output will contain list-cols.
# • Use `values_fn = list` to suppress this warning.
# • Use `values_fn = {summary_fun}` to summarise duplicates.
# • Use the following dplyr code to identify duplicates.
#   {data} |>
#   dplyr::summarise(n = dplyr::n(), .by = c(row_id, cell_id)) |>
#   dplyr::filter(n > 1L)

row_id	1	3	4	2
integer	list	list	list	list
2	[[character]]	[[character]]	[[character]]	[[NULL]]
3	[[character]]	[[character]]	[[character]]	[[character]]
4	[[character]]	[[character]]	[[NULL]]	[[NULL]]
5	[[character]]	[[character]]	[[character]]	[[NULL]]
6	[[character]]	[[character]]	[[character]]	[[character]]
7	[[character]]	[[character]]	[[NULL]]	[[character]]
8	[[character]]	[[character]]	[[NULL]]	[[character]]
9	[[character]]	[[character]]	[[character]]	[[NULL]]
10	[[character]]	[[character]]	[[character]]	[[NULL]]
11	[[character]]	[[character]]	[[character]]	[[NULL]]
12	[[character]]	[[character]]	[[character]]	[[character]]
13	[[character]]	[[NULL]]	[[character]]	[[NULL]]
n: 12

Getting headers requires another operation:

table_cells |>
  filter(is_header) |>
  select(row_id, cell_id, text) |>
  pivot_wider(names_from = cell_id, values_from = text)

row_id integer	1
1 character	Petals
2 character	Internode
3 character	Sepal
4 character	Bract

8.2 Import PowerPoint document

The function pptx_summary() returns the content of a PowerPoint document.

example_pptx <- system.file(package = "officer", "doc_examples/example.pptx")
file.copy(example_pptx, to = "static/office/example.pptx")

# [1] TRUE

static/office/example.pptx

doc <- read_pptx("static/office/example.pptx")
content <- pptx_summary(doc)
head(content)

text	id	content_type	slide_id	row_id	cell_id	col_span	row_span	media_file
character	character	character	integer	integer	integer	integer	integer	character
Title	12	paragraph	1
A table	13	paragraph	1
and some text	13	paragraph	1
and some list (1)	13	paragraph	1
and some list (2)	13	paragraph	1
Header 1	18	table cell	1	1	1	1	1
n: 6

Explore the results:

content |>
  group_by(content_type) |>
  summarise(n = n_distinct(id))

content_type	n
character	integer
image	1
paragraph	5
table cell	2
n: 3

To get all paragraphs:

par_data <- content |>
  filter(content_type %in% "paragraph") |>
  select(id, text)
head(par_data)

id	text
character	character
12	Title
13	A table
13	and some text
13	and some list (1)
13	and some list (2)
15	R logo
n: 6

To get an image:

image_row <- filter(content, content_type %in% "image")
img  <- file.path(tempdir(), "extract.png")
media_extract(doc, path = image_row$media_file, target = img)

# [1] TRUE

8.2.1 PowerPoint tables

Tables are unstacked :

table_cells <- filter(content, content_type %in% "table cell")
head(table_cells)

text	id	content_type	slide_id	row_id	cell_id	col_span	row_span	media_file
character	character	character	integer	integer	integer	integer	integer	character
Header 1	18	table cell	1	1	1	1	1
A	18	table cell	1	2	1	1	1
B	18	table cell	1	3	1	1	1
B	18	table cell	1	4	1	1	1
C	18	table cell	1	5	1	1	1
Header 2	18	table cell	1	1	2	1	1
n: 6

Cells positions and values are dispatched in columns row_id, cell_id, text. Note that here there is no indicator for the table header.

table_cells |>
  filter(id == 18) |>
  select(row_id, cell_id, text) |>
  pivot_wider(names_from = cell_id, values_from = text)

row_id	1	2	3
integer	character	character	character
1	Header 1	Header 2	Header 3
2	A	12.23	blah blah
3	B	1.23	blah blah blah
4	B	9.0	Salut
5	C	6	Hello
n: 5