Split a String By Pattern Matches

These functions split each element of str into substrings. pattern indicates delimiters that separate the input into tokens. The input data between the matches become the fields themselves.

stri_split(str, ..., regex, fixed, coll, charclass)

stri_split_fixed(str, pattern, n = -1L, omit_empty = FALSE, tokens_only = FALSE, simplify = FALSE, ..., opts_fixed = NULL)

stri_split_regex(str, pattern, n = -1L, omit_empty = FALSE, tokens_only = FALSE, simplify = FALSE, ..., opts_regex = NULL)

stri_split_coll(str, pattern, n = -1L, omit_empty = FALSE, tokens_only = FALSE, simplify = FALSE, ..., opts_collator = NULL)

stri_split_charclass(str, pattern, n = -1L, omit_empty = FALSE, tokens_only = FALSE, simplify = FALSE)


character vector with strings to search in


supplementary arguments passed to the underlying functions, including additional settings for opts_collator, opts_regex, opts_fixed, and so on

pattern, regex, fixed, coll, charclass

character vector defining search patterns; for more details refer to stringi-search


integer vector, maximal number of strings to return, and, at the same time, maximal number of text boundaries to look for


logical vector; determines whether empty tokens should be removed from the result (TRUE or FALSE) or replaced with NAs (NA)


single logical value; may affect the result if n is positive, see Details


single logical value; if TRUE or NA, then a character matrix is returned; otherwise (the default), a list of character vectors is given, see Value

opts_collator, opts_fixed, opts_regex

a named list used to tune up a search engine's settings; see stri_opts_collator, stri_opts_fixed, and stri_opts_regex, respectively; NULL for default settings;


Vectorized over str, pattern, n, and omit_empty.

If n is negative, then all pieces are extracted. Otherwise, if tokens_only is FALSE (this is the default, for compatibility with the stringr package), then n-1 tokes are extracted (if possible) and the n-th string gives the remainder (see Examples). On the other hand, if tokens_only is TRUE, then only full tokens (up to n pieces) are extracted.

omit_empty is applied during the split process: if it is set to TRUE, then tokens of zero length are ignored. Thus, empty strings will never appear in the resulting vector. On the other hand, if omit_empty is NA, then empty tokes are substituted with missing strings.

Empty search patterns are not supported. If you would like to split a string into individual characters, use e.g. stri_split_boundaries(str, type="character") for THE Unicode way.

stri_split is a convenience function. It calls either stri_split_regex, stri_split_fixed, stri_split_coll, or stri_split_charclass, depending on the argument used. Relying on one of those underlying functions will make your code run slightly faster.


If simplify=FALSE (the default), then the functions return a list of character vectors.

Otherwise, stri_list2matrix with byrow=TRUE and n_min=n arguments is called on the resulting object. In such a case, a character matrix with an appropriate number of rows (according to the length of str, pattern, etc.) is returned. Note that stri_list2matrix's fill argument is set to an empty string and NA, for simplify equal to TRUE and NA, respectively.

See Also

Other search_split: stri_split_boundaries, stri_split_lines, stringi-search

  • stri_split
  • stri_split_fixed
  • stri_split_regex
  • stri_split_coll
  • stri_split_charclass
stri_split_fixed("a_b_c_d", "_")
stri_split_fixed("a_b_c__d", "_")
stri_split_fixed("a_b_c__d", "_", omit_empty=TRUE)
stri_split_fixed("a_b_c__d", "_", n=2, tokens_only=FALSE) # "a" & remainder
stri_split_fixed("a_b_c__d", "_", n=2, tokens_only=TRUE) # "a" & "b" only
stri_split_fixed("a_b_c__d", "_", n=4, omit_empty=TRUE, tokens_only=TRUE)
stri_split_fixed("a_b_c__d", "_", n=4, omit_empty=FALSE, tokens_only=TRUE)
stri_split_fixed("a_b_c__d", "_", omit_empty=NA)
stri_split_fixed(c("ab_c", "d_ef_g", "h", ""), "_", n=1, tokens_only=TRUE, omit_empty=TRUE)
stri_split_fixed(c("ab_c", "d_ef_g", "h", ""), "_", n=2, tokens_only=TRUE, omit_empty=TRUE)
stri_split_fixed(c("ab_c", "d_ef_g", "h", ""), "_", n=3, tokens_only=TRUE, omit_empty=TRUE)

stri_list2matrix(stri_split_fixed(c("ab,c", "d,ef,g", ",h", ""), ",", omit_empty=TRUE))
stri_split_fixed(c("ab,c", "d,ef,g", ",h", ""), ",", omit_empty=FALSE, simplify=TRUE)
stri_split_fixed(c("ab,c", "d,ef,g", ",h", ""), ",", omit_empty=NA, simplify=TRUE)
stri_split_fixed(c("ab,c", "d,ef,g", ",h", ""), ",", omit_empty=TRUE, simplify=TRUE)
stri_split_fixed(c("ab,c", "d,ef,g", ",h", ""), ",", omit_empty=NA, simplify=NA)

stri_split_regex(c("ab,c", "d,ef  ,  g", ",  h", ""),
   "\\p{WHITE_SPACE}*,\\p{WHITE_SPACE}*", omit_empty=NA, simplify=TRUE)

stri_split_charclass("Lorem ipsum dolor sit amet", "\\p{WHITE_SPACE}")
stri_split_charclass(" Lorem  ipsum dolor", "\\p{WHITE_SPACE}", n=3,
   omit_empty=c(FALSE, TRUE))

stri_split_regex("Lorem ipsum dolor sit amet",
   "\\p{Z}+") # see also stri_split_charclass

# }
Documentation reproduced from package stringi, version 1.1.6, License: file LICENSE

Community examples

Looks like there are no examples yet.