Take the 2-minute tour ×
Unix & Linux Stack Exchange is a question and answer site for users of Linux, FreeBSD and other Un*x-like operating systems.. It's 100% free, no registration required.

I am a Windows user and coming from Windows with all the indexing and eternal wait times when searching for files, I find it surprising how fast does locate work or the autocompletion (that I know) work in linux.

Is there any indexing being done in the background or how is this achieved? I still have a pretty clean installation, so maybe this goes bad with time, but in Windows the moment you want to search on a folder which is not indexed you have to wait for some seconds.


share|improve this question
add comment

2 Answers

up vote 13 down vote accepted

I find it surprising how fast does locate work or the autocompletion (that I know) work in linux. ... Is there any indexing being done in the background or how is this achieved?

This is actually two completely distinct questions.

locate uses an index (slocate stores it in /var/lib/slocate/), that is updated by a nightly cron job. This nightly job typically runs at about 1 or 2AM local time, and completely scans your entire system (including all connected drives). The resulting index is simply a list of filenames.

Auto-complete is handled by your shell. Most systems use bash, so bash-completion is the collection of scripts that manage how this works. (zsh has a similarly-named collection, and most of the other shells have some form of completion built-in.) When Tab is pressed, the shell runs a script that decides, based on what you've typed already, what, exactly, needs to be completed. The script then generates a list of possible completions, which may or may not be the list of files in the current directory, or the list of executable files in your $PATH. The locate command is normally not used for this.

share|improve this answer
Thanks! I see, so it's not so "magical" :) –  oli206 Oct 10 '10 at 18:48
greyfade has ruined the magic forever -_- –  Michael Mrozek Oct 10 '10 at 22:14
@Michael Mrozek: Sorry. :( –  greyfade Oct 11 '10 at 2:30
The shell runs a script sounds not 100% correct, when implementing a shell people shouldn't be misguided in thinking they should write external scripts. It's simply a routine in the code itself, it may or not access external files for configuration convenience, but simple built-in completion works well. –  Aki Apr 20 '12 at 17:14
@Aki: The majority of customized autocompletion in bash is implemented by scripts in /etc/bash_completion.d/. You're right, they need not be shell scripts, but most of the time they are. (zsh keeps its completion scripts in /usr/share/zsh/site-functions/) –  greyfade Apr 20 '12 at 18:27
show 2 more comments

Usually, locate uses an index which is generated once a day via a cron-job (/etc/cron.daily/mlocate on my system for example). It does nothing fancy, basically a complete filesystem traversal with some optimizations and the building of the index data structure.

Shells use probably some internal caching for command completion, but does not use a global index file. Besides, usually, Unix kernels maintain a dentry-cache, i.e.\ they cache file directory information used by directory listings etc. in memory (including the stat'ing of non-existent files - which is also called inverse caching).

share|improve this answer
add comment

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.