yaml (version 2.3.7)

yaml.load: Convert a YAML string into R objects


Parse a YAML string and return R objects.


yaml.load(string, as.named.list = TRUE, handlers = NULL, error.label = NULL,
            eval.expr = getOption("yaml.eval.expr", FALSE),
            merge.precedence = c("order", "override"), merge.warning = FALSE)
  yaml.load_file(input, error.label, readLines.warn=TRUE, ...)


If the root YAML object is a map, a named list or list with an attribute of 'keys' is returned. If the root object is a sequence, a list or vector is returned, depending on the contents of the sequence. A vector of length 1 is returned for single objects.



the YAML string to be parsed


whether or not to return a named list for maps (TRUE by default)


named list of custom handler functions for YAML types (see Details)


a filename or connection; if input is a filename, that file must be encoded in UTF-8


a label to prepend to error messages (see Details)


whether or not to evaluate expressions found in the YAML document (see Details)


behavior of precedence during map merges (see Details)


whether or not to warn about ignored key/value pairs during map merges


logical (default:TRUE) suppress warnings from readLines used inside read_yaml


arguments to pass to yaml.load


Jeremy Stephens <jeremy.f.stephens@vumc.org>


Use yaml.load to load a YAML string. For files and connections, use yaml.load_file, which calls yaml.load with the contents of the specified file or connection.

Sequences of uniform data (e.g. a sequence of integers) are converted into vectors. If the sequence is not uniform, it's returned as a list. Maps are converted into named lists by default, and all the keys in the map are converted to strings. If you don't want the keys to be coerced into strings, set as.named.list to FALSE. When it's FALSE, a list will be returned with an additional attribute named 'keys', which is a list of the un-coerced keys in the map (in the same order as the main list).

You can specify custom handler functions via the handlers argument. This argument must be a named list of functions, where the names are the YAML types (i.e., 'int', 'float', 'seq', etc). The functions you provide will be passed one argument. Custom handler functions for string types (all types except sequence and map) will receive a character vector of length 1. Custom sequence functions will be passed a list of objects. Custom map functions will be passed the object that the internal map handler creates, which is either a named list or a list with a 'keys' attribute (depending on as.named.list). ALL functions you provide must return an object. See the examples for custom handler use.

You can specify a label to be prepended to error messages via the error.label argument. When using yaml.load_file, you can either set the error.label argument explicitly or leave it missing. If missing, yaml.load_file will make an educated guess for the value of error.label by either using the specified filename (when input is a character vector) or using the description of the supplied connection object (via the summary function). You can explicity set error.label to NULL if you don't want to use this functionality.

There is a built-in handler that will evaluate expressions that are tagged with the ‘!expr’ tag. Currently this handler is disabled by default for security reasons. If a ‘!expr’ tag exists and this is set to FALSE a warning will occur. Alternately, you can set the option named ‘yaml.eval.expr’ via the options function to turn on evaluation.

The merge.precedence parameter controls how merge keys are handled. The YAML merge key specification is not specific about how key/value conflicts are resolved during map merges. As a result, various YAML library implementations vary in merge key behavior (notably Python and Ruby). This package's default behavior (when merge.precedence is ‘order’) is to give precedence to key/value pairs that appear first. If you set merge.precedence to ‘override’, natural map key/value pairs will override any duplicate keys found in merged maps, regardless of order. This is the default behavior in Python's YAML library.

This function uses the YAML parser provided by libyaml, which conforms to the YAML 1.1 specification.


YAML: http://yaml.org

libyaml: https://pyyaml.org/wiki/LibYAML

YAML merge specification: http://yaml.org/type/merge.html

See Also



Run this code
  yaml.load("- hey\n- hi\n- hello")
  yaml.load("foo: 123\nbar: 456")
  yaml.load("- foo\n- bar\n- 3.14")
  yaml.load("foo: bar\n123: 456", as.named.list = FALSE)

if (FALSE) {
  # reading from a file (uses readLines internally)
  filename <- tempfile()
  cat("foo: 123", file=filename, sep="\n")

  # custom scalar handler
  my.float.handler <- function(x) { as.numeric(x) + 123 }
  yaml.load("123.456", handlers=list("float#fix"=my.float.handler))

  # custom sequence handler
  yaml.load("- 1\n- 2\n- 3", handlers=list(seq=function(x) { as.integer(x) + 3 }))

  # custom map handler
  yaml.load("foo: 123", handlers=list(map=function(x) { x$foo <- x$foo + 123; x }))

  # handling custom types
  yaml.load("!sqrt 555", handlers=list(sqrt=function(x) { sqrt(as.integer(x)) }))
  yaml.load("!foo\n- 1\n- 2", handlers=list(foo=function(x) { as.integer(x) + 1 }))
  yaml.load("!bar\none: 1\ntwo: 2", handlers=list(foo=function(x) { x$one <- "one"; x }))

  # loading R expressions
  # NOTE: this will not be done by default in the near future
  doc <- yaml.load("inc: !expr function(x) x + 1", eval.expr=TRUE)

  # adding a label to error messages
  try(yaml.load("*", error.label = "foo"))

Run the code above in your browser using DataLab