locate - list files in databases that match a pattern
[-d path | --database=path] [-e | -E | --[non-]existing] [-i |
--ignore-case] [-0 | --null] [-c | --count] [-w | --wholename] [-b |
--basename] [-l N | --limit=N] [-S | --statistics] [-r | --regex ]
[--max-database-age D] [-P | -H | --nofollow] [-L | --follow] [--version] [-A
| --all] [-p | --print] [--help] pattern...
This manual page documents the GNU version of locate
. For each given
searches one or more databases of file names and
displays the file names that contain the pattern. Patterns can contain
shell-style metacharacters: `*', `?', and `'. The metacharacters do not
treat `/' or `.' specially. Therefore, a pattern `foo*bar' can match a file
name that contains `foo3/bar', and a pattern `*duck*' can match a file name
that contains `lake/.ducky'. Patterns that contain metacharacters should be
quoted to protect them from expansion by the shell.
If a pattern is a plain string — it contains no metacharacters —
displays all file names in the database that contain that string
anywhere. If a pattern does contain metacharacters, locate
displays file names that match the pattern exactly. As a result, patterns that
contain metacharacters should usually begin with a `*', and will most often
end with one as well. The exceptions are patterns that are intended to
explicitly match the beginning or end of a file name.
The file name databases contain lists of files that were on the system when the
databases were last updated. The system administrator can choose the file name
of the default database, the frequency with which the databases are updated,
and the directories for which they contain entries; see updatedb(1)
's output is going to a terminal, unusual characters in the
output are escaped in the same way as for the -print action of the find
command. If the output is not going to a terminal, file names are printed
- -0, --null
- Use ASCII NUL as a separator, instead of newline.
- -A, --all
- Print only names which match all non-option arguments, not those matching
one or more non-option arguments.
- -b, --basename
- Results are considered to match if the pattern specified matches the final
component of the name of a file as listed in the database. This final
component is usually referred to as the `base name'.
- -c, --count
- Instead of printing the matched filenames, just print the total number of
matches we found, unless -- print (-p) is also present.
- -d path, --database=path
- Instead of searching the default file name database, search the file name
databases in path, which is a colon-separated list of database file
names. You can also use the environment variable LOCATE_PATH to set
the list of database files to search. The option overrides the environment
variable if both are used. Empty elements in the path are taken to be
synonyms for the file name of the default database. A database can be
supplied on stdin, using `-' as an element of path. If more than
one element of path is `-', later instances are ignored (and a
warning message is printed).
- The file name database format changed starting with GNU find and
locate version 4.0 to allow machines with different byte orderings
to share the databases. This version of locate can automatically
recognize and read databases produced for older versions of GNU
locate or Unix versions of locate or find. Support
for the old locate database format will be discontinued in a future
- -e, --existing
- Only print out such names that currently exist (instead of such names that
existed when the database was created). Note that this may slow down the
program a lot, if there are many matches in the database. If you are using
this option within a program, please note that it is possible for the file
to be deleted after locate has checked that it exists, but before
you use it.
- -E, --non-existing
- Only print out such names that currently do not exist (instead of such
names that existed when the database was created). Note that this may slow
down the program a lot, if there are many matches in the database.
- Print a summary of the options to locate and exit.
- -i, --ignore-case
- Ignore case distinctions in both the pattern and the file names.
- -l N, --limit=N
- Limit the number of matches to N. If a limit is set via this option, the
number of results printed for the -c option will never be larger than this
- -L, --follow
- If testing for the existence of files (with the -e or -E options),
consider broken symbolic links to be non-existing. This is the
- --max-database-age D
- Normally, locate will issue a warning message when it searches a
database which is more than 8 days old. This option changes that value to
something other than 8. The effect of specifying a negative value is
- -m, --mmap
- Accepted but does nothing, for compatibility with BSD locate.
- -P, -H, --nofollow
- If testing for the existence of files (with the -e or -E options), treat
broken symbolic links as if they were existing files. The -H form of this
option is provided purely for similarity with find; the use of -P
is recommended over -H.
- -p, --print
- Print search results when they normally would not, because of the presence
of --statistics (-S) or --count (-c).
- -r, --regex
- The pattern specified on the command line is understood to be a regular
expression, as opposed to a glob pattern. The Regular expressions work in
the same was as in emacs and find, except for the fact that
"." will match a newline. Filenames whose full paths match the
specified regular expression are printed (or, in the case of the -c
option, counted). If you wish to anchor your regular expression at the
ends of the full path name, then as is usual with regular expressions, you
should use the characters ^ and $ to signify this.
- -s, --stdio
- Accepted but does nothing, for compatibility with BSD locate.
- -S, --statistics
- Print various statistics about each locate database and then exit without
performing a search, unless non-option arguments are given. For
compatibility with BSD, -S is accepted as a synonym for --statistics.
However, the output of locate -S is different for the GNU and BSD
implementations of locate.
- Print the version number of locate and exit.
- -w, --wholename
- Match against the whole name of the file as listed in the database. This
is the default.
- Colon-separated list of databases to search. If the value has a leading or
trailing colon, or has two colons in a row, you may get results that vary
between different versions of locate.
The full documentation for locate
is maintained as a Texinfo manual. If
programs are properly installed at your
site, the command info locate
should give you access to the complete
program started life as the BSD fast find program, contributed
to BSD by James A. Woods. This was described by his paper Finding Files
which was published in Usenix ;login:
, Vol 8, No 1,
February/March, 1983, pp. 8-10. When the find
program began to assume a
action if no action was specified, this changed the
interpretation of find pattern
. The BSD developers therefore
moved the fast find functionality into locate
. The GNU implementation
appears to be derived from the same code.
Significant changes to locate
in reverse order:
|| Byte-order independent support for old database format
|| locate -i supports multi-byte characters correctly
|| Introduced --max_db_age
|| Support for the slocate database format
|| Introduced the --all option
|| Introduced the --regex option
|| Introduced options -L, -P, -H
|| Empty items in LOCATE_PATH now indicate the default
|| Introduced the --statistics option
|| Introduced --count and --limit
|| Glob characters cause matching against the whole file name
|| Introduced the LOCATE02 database format
|| Locate can search multiple databases
The locate database correctly handles filenames containing newlines, but only if
the system's sort command has a working -z option. If you suspect that
may need to return filenames containing newlines, consider using
The best way to report a bug is to use the form at
The reason for this is that you
will then be able to track progress in fixing the problem. Other comments
and about the findutils package in general can be sent
to the bug-findutils
mailing list. To join the list, send email to