incadata (version 0.9.1)

best_match: Tries to correct misspelling of character string

Description

This function uses fuzzy string matching to replace one possibly misspelled (or in other way not fully correct) character string with a correct version of the same string.

Usage

best_match(x, key, clean_text = TRUE, no_match = NA, all = FALSE)

Arguments

x

is a character string (or a character vector) that should be matched to the key

key

is a vector containing the correct spellings of the character strings.

clean_text

(boolean of length one) should arguments x and key be passed to clean_text before matched (to ignore special characters)?

no_match

Output value if there is no match. Default is NA. The input is returned unchanged if not matched and no_match = NULL.

all

is a boolean indicator to specify what happens if there is more than one match. Default is FALSE resulting in a warning message and that only the first match is used. The result can then be returned as a vector. If TRUE, all possible matches are returned and the result must therefore be a list.

Value

The function returns a character vector of the same length as x if all = FALSE but with each element substituted to its best match in the key-vector. Strings that could not be matched are NA if (no_match = TRUE) or unchanged if no_match = FALSE. If all = TRUE, one input character string could result in more than one output character string. The output might therefore be longer than the input.

See Also

clean_text

Examples

# NOT RUN {
best_match(c("Hej_apa!", "erik", "babian"), 
  c("hej apa", "hej bepa", "kungen", "Erik"))
best_match(c("Hej_apa", "erik", "babian"),
   c("hej apa", "hej bepa", "kungen", "Erik"), no_match = FALSE)
# }