stringi (version 1.7.8)

stri_locate_all: Locate Pattern Occurrences


These functions find the indexes (positions) where there is a match to some pattern. The functions stri_locate_all_* locate all the matches. stri_locate_first_* and stri_locate_last_* give the first and the last matches, respectively.


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

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

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

stri_locate( str, ..., regex, fixed, coll, charclass, mode = c("first", "all", "last") )

stri_locate_all_charclass( str, pattern, merge = TRUE, omit_no_match = FALSE, get_length = FALSE )

stri_locate_first_charclass(str, pattern, get_length = FALSE)

stri_locate_last_charclass(str, pattern, get_length = FALSE)

stri_locate_all_coll( str, pattern, omit_no_match = FALSE, get_length = FALSE, ..., opts_collator = NULL )

stri_locate_first_coll( str, pattern, get_length = FALSE, ..., opts_collator = NULL )

stri_locate_last_coll( str, pattern, get_length = FALSE, ..., opts_collator = NULL )

stri_locate_all_regex( str, pattern, omit_no_match = FALSE, capture_groups = FALSE, get_length = FALSE, ..., opts_regex = NULL )

stri_locate_first_regex( str, pattern, capture_groups = FALSE, get_length = FALSE, ..., opts_regex = NULL )

stri_locate_last_regex( str, pattern, capture_groups = FALSE, get_length = FALSE, ..., opts_regex = NULL )

stri_locate_all_fixed( str, pattern, omit_no_match = FALSE, get_length = FALSE, ..., opts_fixed = NULL )

stri_locate_first_fixed( str, pattern, get_length = FALSE, ..., opts_fixed = NULL )

stri_locate_last_fixed( str, pattern, get_length = FALSE, ..., opts_fixed = NULL )


For stri_locate_all_*, a list of integer matrices is returned. Each list element represents the results of a separate search scenario. The first column gives the start positions of the matches, and the second column gives the end positions. Moreover, two NAs in a row denote NA arguments or a no-match (the latter only if omit_no_match is FALSE).

stri_locate_first_* and stri_locate_last_*

return an integer matrix with two columns, giving the start and end positions of the first or the last matches, respectively, and two NAs if and only if they are not found.

For stri_locate_*_regex, if the match is of zero length, end will be one character less than start. Note that stri_locate_last_regex searches from start to end, but skips overlapping matches, see the example below.

Setting get_length=TRUE results in the 2nd column representing the length of the match instead of the end position. In this case, negative length denotes a no-match.

If capture_groups=TRUE, then the outputs are equipped with the capture_groups attribute, which is a list of matrices giving the start-end positions of matches to parenthesized subexpressions. Similarly to stri_match_regex, capture group names are extracted unless looking for first/last occurrences of many different patterns.



character vector; strings to search in


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


single string; one of: 'first' (the default), 'all', 'last'

pattern, regex, fixed, coll, charclass

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


single logical value; indicates whether consecutive sequences of indexes in the resulting matrix should be merged; stri_locate_all_charclass only


single logical value; if TRUE, a no-match will be indicated by a matrix with 0 rows stri_locate_all_* only


single logical value; if FALSE (default), generate from-to matrices; otherwise, output from-length ones

opts_collator, opts_fixed, opts_regex

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


single logical value; whether positions of matches to parenthesized subexpressions should be returned too (as capture_groups attribute); stri_locate_*_regex only


Marek Gagolewski and other contributors


Vectorized over str and pattern (with recycling of the elements in the shorter vector if necessary). This allows to, for instance, search for one pattern in each string, search for each pattern in one string, and search for the i-th pattern within the i-th string.

The matches may be extracted by calling stri_sub or stri_sub_all. Alternatively, you may call stri_extract directly.

stri_locate, stri_locate_all, stri_locate_first, and stri_locate_last are convenience functions. They just call stri_locate_*_*, depending on the arguments used.

See Also

The official online manual of stringi at

Gagolewski M., stringi: Fast and portable character string processing in R, Journal of Statistical Software 103(2), 2022, 1-59, tools:::Rd_expr_doi("10.18637/jss.v103.i02")

Other search_locate: about_search, stri_locate_all_boundaries()

Other indexing: stri_locate_all_boundaries(), stri_sub_all(), stri_sub()


Run this code
stri_locate_all('stringi', fixed='i')

stri_locate_first_coll('hladn\u00FD', 'HLADNY', strength=1, locale='sk_SK')

    c('breakfast=eggs;lunch=pizza', 'breakfast=spam', 'no food here'),
)  # named capture groups

stri_locate_all_fixed("abababa", "ABA", case_insensitive=TRUE, overlap=TRUE)
stri_locate_first_fixed("ababa", "aba")
stri_locate_last_fixed("ababa", "aba")  # starts from end
stri_locate_last_regex("ababa", "aba")  # no overlaps, from left to right

x <- c("yes yes", "no", NA)
stri_locate_all_fixed(x, "yes")
stri_locate_all_fixed(x, "yes", omit_no_match=TRUE)
stri_locate_all_fixed(x, "yes", get_length=TRUE)
stri_locate_all_fixed(x, "yes", get_length=TRUE, omit_no_match=TRUE)
stri_locate_first_fixed(x, "yes")
stri_locate_first_fixed(x, "yes", get_length=TRUE)

# Use regex positive-lookahead to locate overlapping pattern matches:
stri_locate_all_regex('ACAGAGACTTTAGATAGAGAAGA', '(?=AGA)')
# note that start > end here (match of length zero)

Run the code above in your browser using DataCamp Workspace