Searching files with locate
If you are feeling lazy and need to search for some file or directory in your whole system, then locate might be your best friend. It can find things super fast from thousands of items with the help of a database created earlier and usually updated automatically once a day.
locate almost works like the librarian in a library. When you ask the librarian for a book, he/she can find it for you in the entire library very efficiently because the books are sorted by some order.
locate does not locate the file/directories on your system directly but on a database of absolute pathnames which are gathered beforehand. This database has an index, which makes finding things very fast.
Using locate is very simple. Let's get started by first installing it.
Installing
In most modern Linux distributions the locate command is available via the plocate and mlocate packages. Both packages gives us two commands: locate and updatedb.
I recommend installing the plocate package because it's faster than mlocate and almost a drop-in-replacement for mlocate. While all variants of locates are more or less the same, in this article I will discuss about plocate particularly.
plocate is available in the official repositores of the following distros and can be installed there like below:
# Arch
sudo pacman -S plocate
# Debian / Ubuntu
sudo apt install plocate
# Fedora
sudo dnf install plocate
Creating the database with updatedb
Before using it you will need to create it's database by running updatedb like below, if you don't want to wait for the system to do it for you:
sudo updatedb
It will take some time. Usually updatedb is run once a day automatically by your system to keep the database updated.
Since the database is not guaranteed to hold the most recent info about your syestem, you will may notice that very recent files/directories do not show up when using locate. To overcome this, you can run the updatedb command manually like above.
How to use the locate command
The general syntax of locate command is:
locate [OPTION]... PATTERN...
Based on the given patterns and options, locate performs a quick database search of absolute pathnames of files and directories that are accessible to the user in the entire system.
It can be used with or without options. First let's see it's usage without any options.
Without options
To find out all the pathnames containing the substring bin we can enter bin as pattern:
locate bin
It will display a long list of absolute pathnames containing the pattern bin in them:
/bin
/sbin
/etc/bindresvport.blacklist
/etc/binfmt.d
/etc/libinput
/etc/ca-certificates/extracted/edk2-cacerts.bin
/etc/profile.d/perlbin.csh
/etc/profile.d/perlbin.sh
.
.
.
Now if we want only the pathnames that contains both bin and zip we can enter both as patterns:
locate bin zip
This time the output will be significantly less somewhat like below:
/usr/bin/bunzip2
/usr/bin/bzip2
/usr/bin/bzip2recover
/usr/bin/funzip
/usr/bin/gunzip
/usr/bin/gzip
/usr/bin/unzip
/usr/bin/unzipsfx
/usr/bin/zipgrep
/usr/bin/zipinfo
We can also use glob patterns(aka wildcard patterns).refs:glob In this case it's important to quote the pattern to prevent the shell from expanding them. For example:
locate '*.md'
It lists only the pathnames that ends with .md. These are the first few items from my system:
/home/ashutosh/.cache/tldr/pages/common/bg.md
/home/ashutosh/.cache/tldr/pages/common/cat.md
/home/ashutosh/.cache/tldr/pages/common/exec.md
/home/ashutosh/.cache/tldr/pages/common/fg.md
/home/ashutosh/.cache/tldr/pages/common/git.md
/home/ashutosh/.cache/tldr/pages/common/neofetch.md
/home/ashutosh/.cache/tldr/pages/common/tldr.md
/home/ashutosh/.cache/tldr/pages/common/ts-node.md
/home/ashutosh/.cache/tldr/pages/common/vim.md
/home/ashutosh/.cache/tldr/pages/linux/pacman.md
Other wildcards that can be used like this are ? for matching any single character and [] for character classes.
With options
Here we will see the most commonly used options:
-c or --count
Adding -c or --count option displays the number of matched items. For example:
locate -c '*.js'
-l or --limit
With the -l or --limit option you can set the upper limit for searching items. For example:
locate -l 10 '*.css'
This will only display at most 10 pathnames.
-i or --ignore-case
By default locate performs case-sensitive search. To do a case-insensitive search add the -i or --ignore-case flag:
locate -i An-Old-Post.md
-e or --existing
After the most recent database update, if some file/directory deleted, it will still show up, if searched for. To not include such items add the -e or --existing flag.
Note that newly added files after the most recent database update will not show up. To show them, the database must be updated.
Options for regular expressions
If you want to search using a POSIX basic regular expression, -r or --regexp are the options. For example to count for all JavaScript and JSON files, the command will be:
locate --regexp -c '\.\(js\|json\)$'
If you want to search using a POSIX extended regular expression, then use the option --regex. So the above command can be written much more cleanly like below with --regex:
locate --regex -c '\.(js|json)$'
Note that these options will treat all given patterns to be the corresponding type of regular expressions.
Summary
# for updating the database
sudo updatedb
# finding by matching a pattern string as substring in the pathnames
locate cat.jpg
# all pathnames will be displayed that satisfy both patterns
locate bin python
# finding using wildcard patterns
locate '*.jpg'
# count all jpg files
locate -c '*.jpg'
# display only first 10 jpg files
locate -l 10 '*.jpg'
# Case-insensitive search
locate -i 'cat.jpg'
# For not including deleted files after most recent database update
locate -e 'cat.jpg'
# POSIX basic regular expression
locate -r '\.\(js\|json\)$'
# POSIX extended regular expression
locate --regex '\.(js|json)$'
For further info see locate's man page by entering man plocate.