How to combine two RMarkdown (.Rmd) files into a single output?

Question

I have two files in the same folder: chapter1.Rmd and chapter2.Rmd, with the following content:

chapter1.Rmd

---
title: "Chapter 1"
output: pdf_document
---

## This is chapter 1. {#Chapter1}

Next up: [chapter 2](#Chapter2)

chapter2.Rmd

---
title: "Chapter 2"
output: pdf_document
---

## This is chapter 2. {#Chapter2}

Previously: [chapter 1](#Chapter1)

How can I knit these so that they combine into a single pdf output?

Of course, render(input = "chapter1.Rmd", output_format = "pdf_document") works perfectly but render(input = "chapter1.Rmd", input = "chapter2.Rmd", output_format = "pdf_document") does not.

Why do I want to do this? To break up a giant document into logical files.

I've used @hadley 's bookdown package to build latex from .Rmd but this seems like overkill for this particular task. Is there a simple solution using knitr/pandoc/linux command line I'm missing? Thanks.

Answer 1

August, 2018 update: This answer was written before the advent of bookdown, which is a more powerful approach to writing Rmarkdown based books. Check out the minimal bookdown example in @Mikey-Harper's answer!

When I want to break a large report into separate Rmd, I usually create a parent Rmd and include the chapters as children. This approach is easy for new users to understand, and if you include a table of contents (toc), it is easy to navigate between chapters.

report.Rmd

---  
title: My Report  
output: 
  pdf_document:
    toc: yes 
---

```{r child = 'chapter1.Rmd'}
```

```{r child = 'chapter2.Rmd'}
```

chapter1.Rmd

# Chapter 1

This is chapter 1.

```{r}
1
```

chapter2.Rmd

# Chapter 2

This is chapter 2.

```{r}
2
```

Build

rmarkdown::render('report.Rmd')

Which produces:

And if you want a quick way to create the chunks for your child documents:

rmd <- list.files(pattern = '*.Rmd', recursive = T)
chunks <- paste0("```{r child = '", rmd, "'}\n```\n")
cat(chunks, sep = '\n')
# ```{r child = 'chapter1.Rmd'}
# ```
#
# ```{r child = 'chapter2.Rmd'}
# ```

Answer 2

I'd recommend that people use the bookdown package for creating reports from multiple R Markdown files. It adds a lot of useful features like cross-referencing which are very useful for longer documents.

Adapting the example from @Eric, here is a minimal example of the bookdown setup. The main detail is that the main file has to be called index.Rmd , and must include the additional YAML line site: bookdown::bookdown_site:

index.Rmd

---
title: "A Minimal bookdown document"
site: bookdown::bookdown_site
output:
  bookdown::pdf_document2:
    toc: yes
---

01-intro.Rmd:

# Chapter 1

This is chapter 1.

```{r}
1
```

02-intro.Rmd:

# Chapter 2

This is chapter 2.

```{r}
2
```

If we Knit the index.Rmd bookdown will merge all the files in the same directory in alphabetical order (this behaviour can be changed using an extra _bookdown.yml file).

Once you get comfortable with this basic setup, it is easy to customise the bookdown document and output formats using additional configuration files i.e. _bookdown.yml and _output.yml

Further Reading

• R Markdown: The definitive Guide: Chapter 11 provides a great overview of bookdown
• Authoring books with bookdown provides a comprehensive guide on bookdown, and recommended for more advanced details.

Answer 3

This worked for me:

Rmd_bind <- 
    function(dir = ".",
    book_header = readLines(textConnection("---\ntitle: 'Title'\n---")))
{
    old <- setwd(dir)
    if(length(grep("book.Rmd", list.files())) > 0){
    warning("book.Rmd already exists")
    }
    write(book_header, file = "book.Rmd", )
    cfiles <- list.files(pattern = "*.Rmd", )
    ttext <- NULL
    for(i in 1:length(cfiles)){
    text <- readLines(cfiles[i])
    hspan <- grep("---", text)
    text <- text[-c(hspan[1]:hspan[2])]
    write(text, sep = "\n", file = "book.Rmd", append = T)
    }
    render("book.Rmd", output_format = "pdf_document")
    setwd(old)
    }

Imagine there's a better solution and would be nice to have something like this in rmarkdown or knitr packages.