janitor

Clean names

clean_names() renames every column of a table, and make_clean_names() applies the same rules to a list of strings. The rules follow janitor’s defaults:

• % becomes percent, # becomes number, and quotes are dropped.
• Accented and non-Latin characters are transliterated to ASCII: Café becomes cafe, straße becomes strasse, and αβγ becomes abg.
• Spaces and punctuation become single underscores, and leading and trailing ones are dropped.
• Case changes split words: camelCaseVar becomes camel_case_var, and HTMLParser becomes html_parser. Digits stay attached to their word.
• A name that would start with a digit gets an x prefix, and an empty name becomes x.
• A repeated name gets _2, _3, and so on, in order of appearance.

Transliteration uses ext-intl, which gives the same results as janitor’s stringi. Without ext-intl, phplyr falls back to iconv, and some names can differ from janitor’s.

make_clean_names(['First Name', 'first_name', '% Done', '1st Place']);
// ['first_name', 'first_name_2', 'percent_done', 'x1st_place']

pipe($survey, clean_names());

Remove empty rows and columns

remove_empty() drops rows and columns that are entirely NA. Pass 'rows' or 'cols' to drop only one kind. With cutoff below 1, a row or column is kept when its share of non-NA values is above the cutoff. A cutoff works with only one of 'rows' and 'cols'. An empty string is a value, as in R.

Duplicates

get_dupes('email') returns the rows whose values in the given columns appear more than once. The result has the given columns first, then dupe_count, then the other columns. Rows are sorted by dupe_count (descending) and then by the given columns. With no columns, every column is used. A grouped table keeps its groups, but duplicates are found across the whole table.

Frequency tables

tabyl('g') counts the values of one column. The result has the values, n, and percent. When the column has NAs, valid_percent gives each share with NA left out. showNa: false drops the NA row instead.

tabyl('g', 'h') counts pairs of values. It has one row per value of g and one column of counts per value of h. An NA in h is counted in the NA_ column, and an empty string in the emptystring_ column.

Values are sorted as in dplyr’s count(), with NA last. Factors keep their level order, and showMissingLevels: false leaves out unused levels.

pipe($people, tabyl('region'));
pipe($people, tabyl('region', 'smoker', showNa: false));

janitor’s adorn_*() functions and three-way tables are not included.