Tag Archives: Word

Exporting R output to MS-Word with R2wd (an example session)

Creating reports is one of the basic tasks in data analysis. R provides numerous functions and packages to export it’s (beautiful) output and help compile it into a report.

In this post I will present one such (basic) solution for Windows OS users for exporting R output into Microsoft Word using the R2wd (package). There are more ways and strategies for doing this, and if encouraged by comments, I will gladly write more on the subject.
* * *

R to Word using {R2wd}

The package R2wd (available through CRAN) relies on rcom. It is a wrapper that uses the statconnDCOM server to communicate with MS-Word via the COM interface.

R2wd can perform the basic tasks you would expect to need when creating a report from R. It allows you to:

  • Create a new Word file
  • Create headers and sub-headers
  • Move to a new pages in the document
  • Write text
  • Insert tables (that is “data.frame” and “matrix”objects)
  • Insert plots
  • Save and close the Word document
  • …(and more)

The current R2wd can still be seen as being in BETA stages.  Some features are not yet available, such as:

  • Choosing text font (which means most of us will need to manually change the font in the document to “couriers new…”, in order for the formatting to look good)
  • Inserting of complex object outputs (such as summery.lm, although in the example bellow I show how that can be achieved using a simple function)
  • Speed – the speed of inserting a table is somewhat slow, I am not sure how it would scale to large documents

But from a (pleasant) correspondence with the package developer, I was assured the next release will supply us with more options and features.

R2wd package developer, Christan Ritter, invites feedback from users.  So if you have features you are missing in this packages, I believe he would like to know about it (you can e-mail Christan at:     christian.ritter <-at-> ridaco <-dot-> be  )

Getting R2wd 1.3

The current version of R2wd is 1.1 and Christan Ritter (the package developer), says it is a “first idea” and that a more elaborate version will soon (e.g: around July) be available on CRAN.   In the meantime, Christan was so kind as to send me a more recent version of the package, which you (until it gets uploaded to CRAN), you are welcome to download from here:
R2wd 1.3 download link

How to use R2wd to create a report – a sample session

Being young doesn’t prevent from R2wd to do some nice things.

Here is the text from the library(help=R2wd) :

If Word is not already running, wdGet() opens a new Word document, otherwise, it establishes a COM handle to the instance which is already running. The functions wdTitle, wdHeader, wdBody, and wdParagraph can be used to inject text elements into Word. Moreover, bookmarks can be added via wdInsertBookmarks and wdGoToBookmark allows to navigate among the bookmarks which also exist. There is another set of convenience functions, wdSection, wdSubsection, and wdSubsubsection which insert headers of level 1, 2, or 3, start new ’Sections’ in Word, and add bookmarks.
Graphs and dataframes can be inserted intoWord, by the wdPlot, wdTable commands. The wdTable command takes a dataframe or an array as arguments, creates a Word table of the appropriate dimensions and injects the content of the dataframe or array into it. It then formats the table in Word using elementary formating elements.
The functions wdApplyTheme and wdApplyTemplate allow to work with themes and templates.

Here is an example sessions to demonstrate some of what is said:

 
# install.packages("R2wd")
# library(help=R2wd)
require(R2wd)
 
 
wdGet(T)	# If no word file is open, it will start a new one - can set if to have the file visiable or not
wdNewDoc("c:\\This.doc")	# this creates a new file with "this.doc" name
 
wdApplyTemplate("c:\\This.dot")	# this applies a template
 
 
wdTitle("Examples of R2wd (a package to write Word documents from R)")	# adds a title to the file
 
wdSection("Example 1 - adding text", newpage = T) # This can also create a header
 
wdHeading(level = 2, "Header 2")
wdBody("This is the first example we will show")
wdBody("(Notice how, by using two different lines in wdBody, we got two different paragraphs)")
wdBody("(Notice how I can use this: '\ n' (without the space), to  \n  go to the next 
		line)")
wdBody("האם זה עובד בעברית ?")
wdBody("It doesn't work with Hebrew...")
wdBody("O.k, let's move to the next page (and the next example)")
 
wdSection("Example 2 - adding tables", newpage = T)
wdBody("Table using 'format'")
wdTable(format(head(mtcars)))
wdBody("Table without using 'format'")
wdTable(head(mtcars))
 
 
wdSection("Example 3 - adding lm summary", newpage = T)
 
## Example from  ?lm 
ctl <- c(4.17,5.58,5.18,6.11,4.50,4.61,5.17,4.53,5.33,5.14)
trt <- c(4.81,4.17,4.41,3.59,5.87,3.83,6.03,4.89,4.32,4.69)
group <- gl(2,10,20, labels=c("Ctl","Trt"))
weight <- c(ctl, trt)
 
# This wouldn't work!
# temp <- summary(lm(weight ~ group))
# wdBody(temp)
 
# Here is a solution for how to implent the summary.lm output to word
wdBody.anything <- function(output)
{
	# This function takes the output of an object and prints it line by line into the word document
	# Notice that in many cases you will need to change the text font into courier new roman...
	a <- capture.output(output)
	for(i in seq_along(a))
	{
		wdBody(format(a[i]))
	}
}
 
temp <- summary(lm(weight ~ group))
wdBody.anything(temp)
 
 
 
wdSection("Example 4 - Inserting some plots", newpage = T)
 
wdPlot(rnorm(100), plotfun = plot, height = 10, width =20, pointsize = 20)
wdPlot(rnorm(100), plotfun = plot, height = 10, width =20, pointsize = 20)
wdPlot(rnorm(100), plotfun = plot, height = 10, width =20, pointsize = 50)
 
# wdPageBreak()
 
 
wdSave("c:\\This.doc") # save current file (can say what file name to use)
wdQuit() # close the word file

Update:
Upon reading my post, Chris suggested that I’ll also add a note here about SWORD, a tool written by Thomas Baier (the creator of the StatconnDCOM server) which allows to include R-code in a Sweave-like fashion in Word documents. Here is a link to the project: http://rcom.univie.ac.at