R YAML package

The R YAML package implements the libyaml YAML parser and emitter for R.

You can see the development history of this package here.

What is YAML?

YAML is a human-readable markup language. With it, you can create easily readable documents that can be consumed by a variety of programming languages.


Hash of baseball teams per league:
  - Boston Red Sox
  - Detroit Tigers
  - New York Yankees
  - New York Mets
  - Chicago Cubs
  - Atlanta Braves

Data dictionary specification:
- field: ID
  description: primary identifier
  type: integer
  primary key: yes
- field: DOB
  description: date of birth
  type: date
  format: yyyy-mm-dd
- field: State
  description: state of residence
  type: string



You can install this package directly from CRAN by running (from within R): install.packages('yaml')


  1. Download the appropriate zip file or tar.gz file from Github: https://github.com/viking/r-yaml/downloads
  2. Run R CMD INSTALL followed by the name of the file you downloaded (as root if necessary)


  1. Download the source via git: git clone git://github.com/viking/r-yaml yaml
  2. Run R CMD check yaml to make sure everything is OK.
  3. Run R CMD INSTALL yaml (as root if necessary).


The yaml packages has two main functions: yaml.load and as.yaml.


The yaml.load function is the YAML parsing function. It accepts a YAML document as a string. Here's a simple example that parses a YAML sequence:
x <- "
- 1
- 2
- 3
yaml.load(x)  #=> [1] 1 2 3


A YAML scalar is the basic building block of YAML documents. Example of a YAML document with one element:

In this case, the scalar "1.2345" is typed as a float (or numeric) by the parser. yaml.load would return a numeric vector of length 1 for this document.
yaml.load("1.2345")  #=> [1] 1.2345


A YAML sequence is a list of elements. Here's an example of a simple YAML sequence:
- this
- is
- a
- simple
- sequence
- of
- scalars

If you pass a YAML sequence to yaml.load, a couple of things can happen. If all of the elements in the sequence are uniform, yaml.load will return a vector of that type (i.e. character, integer, real, or logical). If the elements are not uniform, yaml.load will return a list of the elements. No coercion is done by default.


A YAML map is a list of paired keys and values, or hash, of elements. Here's an example of a simple YAML map:
one: 1
two: 2
three: 3
four: 4

Passing a map to yaml.load will produce a named list by default. That is, keys are coerced to strings. Since it is possible for the keys of a YAML map to be almost anything (not just strings), you might not want yaml.load to return a named list. If you want to preserve the data type of keys, you can pass as.named.list = FALSE to yaml.load. If as.named.list is FALSE, yaml.load will create a keys attribute for the list it returns instead of coercing the keys into strings.


yaml.load has the capability to accept custom handler functions. With handlers, you can customize yaml.load to do almost anything you want. Example of handler usage:

integer.handler <- function(x) { as.integer(x) + 123 }
yaml.load("123", handlers = list(int = integer.handler))  #=> [1] 246

Handlers are passed to yaml.load through the handlers argument. The handlers argument must be a named list of functions, where each name is the YAML type that you want to be handled by your function. The functions you provide must accept one argument and must return an R object.

Handler functions will be passed a string or list, depending on the original type of the object. In the example above, integer.handler was passed the string "123".

Sequence handlers

Custom sequence handlers will be passed a list of objects. You can then convert the list into whatever you want and return it. Example:

sequence.handler <- function(x) {
  tmp <- as.numeric(x)
  tmp / 5
string <- "
- foo
- bar
- 123
- 4.567
yaml.load(string, handlers = list(seq = sequence.handler))  #=> [1]      NA      NA 24.6000  0.9134

Map handlers

Custom map handlers work much in the same way as custom list handlers. A map handler function is passed a named list, or a list with a keys attribute (depending on the value of as.named.list). Example:

string <- "
  - 1
  - 2
  - 3
  - 4
yaml.load(string, handlers = list(map = function(x) { as.data.frame(x) }))

  b a
1 3 1
2 4 2

An interesting thing to note in this example is that the b column appears before the a column in the resulting data frame. This is because YAML maps are considered to be hashes, and therefore, order is not preserved. If you want an ordered map, you can use a combination of maps and sequences like so:

- a:
    - 1
    - 2
- b:
    - 3
    - 4


yaml.load_file does the same thing as yaml.load, except it reads a file from a connection. For example:
x <- yaml.load_file("Data/document.yml")

This function takes the same arguments as yaml.load, with the exception that the first argument is a filename or a connection.


as.yaml is an S3 method used to convert R objects into YAML strings. Example as.yaml usage:
x <- as.yaml(1:5)
cat(x, "\n")

Output from above example:
- 1
- 2
- 3
- 4
- 5


Here's the list of as.yaml arguments:
Name Description Default
x the object to convert  
line.sep line separator to use ("\n", "\r\n", or "\r") "\n"
indent number of spaces to use for indenting 2
column.major determines if data.frames are output as column major TRUE


The column.major option determines how a data frame is converted into YAML. By default, column.major is TRUE.

Example of as.yaml when column.major is TRUE:
x <- data.frame(a=1:5, b=6:10)
y <- as.yaml(x, column.major = TRUE)
cat(y, "\n")

  - 1
  - 2
  - 3
  - 4
  - 5
  - 6
  - 7
  - 8
  - 9
  - 10

x <- data.frame(a=1:5, b=6:10)
y <- as.yaml(x, column.major = FALSE)
cat(y, "\n")

- a: 1
  b: 6
- a: 2
  b: 7
- a: 3
  b: 8
- a: 4
  b: 9
- a: 5
  b: 10

Additional documentation

For more information, run help(package='yaml') or example('yaml-package') for some examples.
Topic revision: r23 - 20 Jan 2012, JeremyStephens

This site is powered by FoswikiCopyright © 2013-2022 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding Vanderbilt Biostatistics Wiki? Send feedback