clean_names: Cleans names of an object (usually a data.frame).

View source: R/clean_names.R

clean_namesR Documentation

Cleans names of an object (usually a data.frame).


Resulting names are unique and consist only of the _ character, numbers, and letters. Capitalization preferences can be specified using the case parameter.

Accented characters are transliterated to ASCII. For example, an "o" with a German umlaut over it becomes "o", and the Spanish character "enye" becomes "n".

This function takes and returns a data.frame, for ease of piping with `%>%`. For the underlying function that works on a character vector of names, see make_clean_names. clean_names relies on the versatile function to_any_case, which accepts many arguments. See that function's documentation for ideas on getting the most out of clean_names. A few examples are included below.

A common issue is that the micro/mu symbol is replaced by "m" instead of "u". The replacement with "m" is more correct when doing Greek-to-ASCII transliteration but less correct when doing scientific data-to-ASCII transliteration. A warning will be generated if the "m" replacement occurs. To replace with "u", please add the argument replace=janitor:::mu_to_u which is a character vector mapping all known mu or micro Unicode code points (characters) to "u".


clean_names(dat, ...)

## Default S3 method:
clean_names(dat, ...)

## S3 method for class 'sf'
clean_names(dat, ...)

## S3 method for class 'tbl_graph'
clean_names(dat, ...)

## S3 method for class 'tbl_lazy'
clean_names(dat, ...)



the input data.frame.


Arguments passed on to make_clean_names


The desired target case (default is "snake") will be passed to snakecase::to_any_case() with the exception of "old_janitor", which exists only to support legacy code (it preserves the behavior of clean_names() prior to addition of the "case" argument (janitor versions <= 0.3.1). "old_janitor" is not intended for new code. See to_any_case for a wide variety of supported cases, including "sentence" and "title" case.


A named character vector where the name is replaced by the value.


Convert the names to ASCII (TRUE, default) or not (FALSE).


Should make.names() be applied to ensure that the output is usable as a name without quoting? (Avoiding make.names() ensures that the output is locale-independent but quoting may be required.)


Allow duplicates in the returned names (TRUE) or not (FALSE, the default).


(short for separator input) if character, is interpreted as a regular expression (wrapped internally into stringr::regex()). The default value is a regular expression that matches any sequence of non-alphanumeric values. All matches will be replaced by underscores (additionally to "_" and " ", for which this is always true, even if NULL is supplied). These underscores are used internally to split the strings into substrings and specify the word boundaries.


An integer that will determine the parsing_option.

  • 1: "RRRStudio" -> "RRR_Studio"

  • 2: "RRRStudio" -> "RRRS_tudio"

  • 3: "RRRStudio" -> "RRRSStudio". This will become for example "Rrrstudio" when we convert to lower camel case.

  • -1, -2, -3: These parsing_options's will suppress the conversion after non-alphanumeric values.

  • 0: no parsing


A character vector (if not NULL). The entries of this argument need to be elements of stringi::stri_trans_list() (like "Latin-ASCII", which is often useful) or names of lookup tables (currently only "german" is supported). In the order of the entries the letters of the input string will be transliterated via stringi::stri_trans_general() or replaced via the matches of the lookup table. When named character elements are supplied as part of 'transliterations', anything that matches the names is replaced by the corresponding value. You should use this feature with care in case of case = "parsed", case = "internal_parsing" and case = "none", since for upper case letters, which have transliterations/replacements of length 2, the second letter will be transliterated to lowercase, for example Oe, Ae, Ss, which might not always be what is intended. In this case you can make usage of the option to supply named elements and specify the transliterations yourself.


A character specifying the alignment of numerals ("middle", left, right, asis or tight). I.e. numerals = "left" ensures that no output separator is in front of a digit.


clean_names() is intended to be used on data.frames and data.frame-like objects. For this reason there are methods to support using clean_names() on sf and tbl_graph (from tidygraph) objects as well as on database connections through dbplyr. For cleaning other named objects like named lists and vectors, use make_clean_names().


Returns the data.frame with clean names.

See Also

Other Set names: find_header(), mu_to_u, row_to_names()


# --- Simple Usage ---
x <- data.frame(caseID = 1, DOB = 2, Other = 3)

# or pipe in the input data.frame:
x %>%

# if you prefer camelCase variable names:
x %>%
  clean_names(., "lower_camel")

# (not run) run clean_names after reading in a spreadsheet:
# library(readxl)
# read_excel("messy_excel_file.xlsx") %>%
#   clean_names()

# --- Taking advantage of the underlying snakecase::to_any_case arguments ---

# Restore column names to Title Case, e.g., for plotting
mtcars %>%
  clean_names(case = "title")
# Tell clean_names to leave certain abbreviations untouched:
x %>%
  clean_names(case = "upper_camel", abbreviations = c("ID", "DOB")) 

janitor documentation built on Feb. 16, 2023, 10:16 p.m.