Currently I'm trying to migrate from mlocate to plocate. I'm using mlocate in my script (note: locate is aliased to either mlocate or plocate depending which one I've installed). ## Successfully searched home directory with mlocate Below is the first 10 outputs after running locate home (locate is aliased to mlocate):

/home
/etc/apparmor.d/tunables/home
/etc/apparmor.d/tunables/home.d
/etc/apparmor.d/tunables/home.d/site.local
/etc/systemd/homed.conf
/home/username
/home/username/.bash_history
/home/username/.bash_profile
/home/username/.bashrc
/home/username/.cache
/home/username/.cargo

As you can see, I could successfully find files in my home directories with mlocate. ## Unsuccessfully searched home directory with plocate: However, after installing plocate, I get results from /etc/, /usr/ etc, and all I get is one /home: After running locate home (locate is aliased to plocate):

/home
/etc/apparmor.d/tunables/home
/etc/apparmor.d/tunables/home.d
/etc/apparmor.d/tunables/home.d/site.local
/etc/systemd/homed.conf
/usr/bin/addgnupghome
/usr/bin/homectl

As you can see, plocate couldn't find files and directories in my home directory. ## What I've tried ### 1. Comment from author of plocate This manjaro thread [How to use plocate?](https://forum.manjaro.org/t/how-to-use-plocate/82592) has the author of plocate commented as below: > First, check that the database has been updated recently. Most users will want to use plocate’s updatedb; plocate-build (which converts from mlocate’s database) is generally not what you want since plocate 1.1.0. [...] > > The other reason why a file isn’t shown, is typically permissions. Check if you can find the files as root (sudo plocate test); if you can, the problem is most likely that you don’t have access rights to the directory all the way down from the root. plocate should find anything that find / -name *test* does, but no more. I've run sudo updatedb. My home dir access right:

/ 
➜ ll
drwxr-xr-x - root 17 Jul  2022 home/
drwxr-xr-x - root 16 Apr 21:44 usr/

/home 
➜ ll
drwxr-xr-x - username 16 Apr 21:51 username/

It does seem that plocate couldn't access files and directories under my home dir, but it has the same access right as usr. I don't understand why plocate can see usr but not files and dir under home. #### 2. Results are different when running as sudo The results are different when I run plocate as sudo. Below is regular locate:

➜  locate ranger
/usr/bin/ranger
/usr/lib/python3.10/site-packages/ranger
/usr/lib/python3.10/site-packages/ranger_fm-1.9.3-py3.10.egg-info
/usr/lib/python3.10/site-packages/ranger/__init__.py
/usr/lib/python3.10/site-packages/ranger/__pycache__

Below run as sudo:

➜  sudo locate ranger
/root/.config/ranger
/root/.local/share/ranger
/root/.local/share/ranger/bookmarks
/root/.local/share/ranger/history
/root/.local/share/ranger/tagged
/usr/bin/ranger
/usr/lib/python3.10/site-packages/ranger
/usr/lib/python3.10/site-packages/ranger_fm-1.9.3-py3.10.egg-info

I can see results in my home directories (i.e., first 5 results). ## What I want I'd like to be able to search my home directory with plocate the way I can do so with mlocate. In other words, I expect results from home directory when using plocate:

$ locate home
/home
/etc/apparmor.d/tunables/home
/etc/apparmor.d/tunables/home.d
/etc/apparmor.d/tunables/home.d/site.local
/etc/systemd/homed.conf
/home/username
/home/username/.bash_history
/home/username/.bash_profile
/home/username/.bashrc
/home/username/.cache
/home/username/.cargo

This 4.20.3 Arch system its BTRFS formatted / disk is without free space left. It turns out that mlocate is the cause: # du -h --exclude=Volumes -- * 2>/dev/null | sort -hr | head -2 11G var 9.6G var/lib/mlocate The question https://unix.stackexchange.com/questions/82985/updatedb-can-not-open-a-temporary-file-for-var-lib-mlocate-mlocate-db accepted answer suggests to prepend sudo, though that doesn't change a thing: # sudo updatedb updatedb: `/var/lib/mlocate/mlocate.db' is locked (probably by an earlier updatedb) There seems to be a temporary file in /var/lib/mlocate that is eating up all disk space: # ls -lh var/lib/mlocate/ -rw-r----- 1 root locate 1.1M Oct 21 00:00 mlocate.db -rw------- 1 root root 9.6G Dec 30 19:46 mlocate.db.PRvfsw Could the root cause be the .timer update job that is hanging? # systemctl status updatedb.timer * updatedb.timer - Daily locate database update Loaded: loaded (/usr/lib/systemd/system/updatedb.timer; static; vendor preset: disabled) Active: active (running) since Mon 2019-10-21 16:05:10 CEST; 2 months 9 days ago Trigger: n/a Both restart and stop don't remove the temporary large .db file and updatedb still returns locked. There seems to be an updatedb process still running: # ps -ef | grep updatedb root 3249 1 99 Oct22 ? 213573-14:47:11 /usr/bin/updatedb I know I can kill this process. The root cause is most likely a faulty USB stick: # ls /Volumes/RM_GUE__ ls: cannot access '/Volumes/RM_GUE__/'$'\001\020': Input/output error ls: cannot access '/Volumes/RM_GUE__/)': Input/output error Though the next time a USB stick becomes faulty, / will fill up again. ###updatedb.conf The updatedb.conf options don't bring me any useful filter option: - by path: I can't guess the name that the partition will have after corruption - by filesystem: In this case VFAT was corrupted (and read-only), though I can't look in to the future which file system will get corrupted. How to resolve this issue neat and permanently, for example by limiting the duration that updatedb.timer may run and/or skipping disks which suffer input/output errors, or limit the file size LimitFSIZE= or even something better?

locate gtags would find all the files named gtags. What if I only need executables, is there any way to do this?

The locate program of findutils scans one or more databases of filenames and displays any matches. This can be used as a very fast find command if the file was present during the last file name database update. There are many kinds of databases nowadays, - relational databases (with query language e.g. SQL), - NoSQL databases - document-oriented databases (e.g. MongoDB) - Key-value database (e.g. Redis) - Column-oriented databases (e.g. Cassandra) - Graph database So what kind of database does updatedb update and locate use?

On a standard desktop installation of Ubuntu 24.04, when I try to use the locate command, searches for files that I clearly have in my home directory do not show up -- it appears as if the locate database has never been updated. However, I can run sudo updatedb, which manually updates the locate database, and I can then search for files as expected. This continues to work for a day, after which searches for files in my home directory no longer work again. So my assumption is that after the locate script in cron.daily runs, the database is altered or erased. One solution would be to just remove the locate script from cron.daily, which I think would "solve" the problem. But is there a more proper way to handle this?

I have read that to run the **locate** command on **CYGWIN**, first I need to run the following command: updatedb --localpaths=c:/ What does this command do exactly? As I run this on CYGWIN and it is taking too much time.

Since locate doesn't work by default in Ubuntu 22.04.1 LTS, and the message on the terminal says:

Command 'locate' not found, but can be installed with:
sudo apt install plocate

As we can see, it says that we can install it with sudo apt install plocate. I want to know the difference between mlocate and plocate so that I can decide which one to install.

The locate command doesn't list any search results from files that are in my home folder. Is there a config file to tweak to enable this behavior? sudo locate word didn't work, nor did updating the locate database. MacBook Pro, Sequoia, bash 3.2.57. Thanks!

When working with the output of commands such as locate which produce lists of paths in "human readable form" (i.e. without \ in front of spaces), how do you redirect their output to another command? The output of $ locate [something] produces paths with spaces in them, which inhibits other programs to use the paths in the case they contain spaces. For example, if I were to $ du -h locate *.doc this will produce an error on all files and directories that contain spaces. (wrapping the ticks in spaces does not work)

When I use the locate command (specifically plocate 1.1.22), it doesn't seem to include any files under the /mnt directory, i.e. on external drives. The man page for updatedb claims "The whole file system is scanned by default" (which I take to mean /**/*), so I retried after an update, but no luck. I considered creating a symlink to /mnt in a place that updatedb *does* map, but I understand from [this Q&A](https://unix.stackexchange.com/a/650592/610644) that symlinks won't work by default. I also tried setting PRUNE_BIND_MOUNTS = "no" and removing /mnt from PRUNEPATHS in /etc/updatedb.conf, but again this did nothing. So, is there a way to make updatedb or locate work with files under /mnt?

I am trying to pass filenames retrieved from everything voidtools via their es.exe command line invoked in wsl environment . Effectively, it works similarly to mlocate. es.exe communicates with everything background process and returns paths that I'm converting with this function to WSL paths

es() {
    # Invoke the es.exe with provided arguments and pipe its output
    # through sed to replace Windows-style line endings with Unix-style line endings,
    # then use xargs to handle newline-delimited input and pass each path as a single argument to wslpath.
    # -p argument - search paths
    # -sort date-modified - sensible default
    # 2>/dev/null - remove "xargs: wslpath: terminated by signal 13"
    /mnt/c/Users/user/Downloads/ES-1.1.0.27.x64/es.exe -p -sort date-modified -instance 1.5a "$@" | sed 's/\r$//' | xargs -n1 -d'\n' wslpath 2>/dev/null
}

this for example can return all file paths that contain "plan" and "png" in their path. Results are newline delimited es plan png results in:

/mnt/long path with spaces/plan1.png
/mnt/long plan with spaces/path2.png
...

I'm trying then to pass it to gwenview. I couldn't pipe it, so I'm trying to pass them as command line arguments.

gwenview "$(es plan png | awk '{printf "\"%s\" ", $0}'))"

I expected it this to exand like this

gwenview "/mnt/long path with spaces/plan1.png" "/mnt/long plan with spaces/path2.png"

because writing it like this (all names copied inline) works. However, trying to do it with process substitution doesn't work, no images are displayed. Gwenview isn't clear what error is unfortunately. My assumption is that it has to do with spaces in the paths and the way they are handled in process substitution. Any ideas would be very appreciated.

Running locate command with a file name returns multiple entries as, a.py a.pyc a.py~ Obviously, i am not looking for binaries/compiled files. I tried to configure locate command to ignore files that has pattern *.pyc, *.py~. Man page shows, locate command doesn't have globbing option. How can one configure locate command to ignore files that match a pattern?

I have created mlocate database with the contents of a particular folder. I see that the updatedb doesn't include the path pointed by the symbolic links in the database. How can I include the path pointed by the symbolic links in the database? **Surprisingly**: *mlocate* has a default option -L or --follow that follows trailing symbolic links when checking file existence (default). > What purpose does it serves when *updatedb* doesn't include symlinks! --- References: - [updatedb(8): update database for mlocate - Linux man page](https://linux.die.net/man/8/updatedb) - [mlocate - Gentoo Wiki](https://wiki.gentoo.org/wiki/Mlocate)

I tried to update the locate database, and for some reason it has been corrupted:

locate foo

locate: locate database header corrupt, bigram char outside 0, 32-127: -1

How to repair it? Or maybe I need to just update it to make it work? And finally, back to the original problem, *how* to update it? (Yes, there are many answers about this, but hey, I tried only the most upvoted ones, and the database has been corrupted. It seems the answers were outdated or macOS has some issues here: most answers were for Linux.) Sonoma 14.4.1, zsh 5.9 (x86_64-apple-darwin23.0)

A search with locate finds paths in the filesystem. Often, you know a priori you are interested in either files only, or directories only. A 'locate' search often returns many results. It would be useful to include only one of the types in the result, because it helps shorten the output. But there is a more interesting argument to leave out either files or directories: because the list of result paths can be ambiguous - not only in theory. The example below is a real world case, and not unusual: $ locate --regex --basename "xfce4-keyboard-overlay$" /usr/local/bin/xfce4-keyboard-overlay /usr/local/share/xfce4-keyboard-overlay Ok, we found something! But... files, or directories? $ file /usr/local/bin/xfce4-keyboard-overlay /usr/local/bin/xfce4-keyboard-overlay: bash script So that is a file... $ file /usr/local/share/xfce4-keyboard-overlay /usr/local/share/xfce4-keyboard-overlay: directory while the second is not. This ambiguity is making long lists of paths hard to read, so it would be really nice to filter directories out, for example using a comman line option for locate. Does something like this exist? Even if the filter for directories is separate from locate? At least, one could use a script to iterate all the file names to check - which may be slow.

I keep my digital music and digital photos in directories in a Windows partition, mounted at /media/win_c on my dual-boot box. I'd like to include those directories—but only those directories—in the locate database. However, as far as I can make out, updatedb.conf only offers options to exclude directories, not add them. Of course, I could remove /media from PRUNEPATHS, and then add a whole bunch of subdirectories (/media/win_c/Drivers, /media/win_c/ProgramData...) but this seems a very clunky way of doing it—surely there's a more elegant solution? (I tried just creating soft links to the Windows directories from an indexed linux partition, but that doesn't seem to help.)

When I don't remember where a file or a folder is, I sometime use the locate command (that finds more occurrences, allow more candidates than a find, in my mind. But maybe I'm mistaking). But then there's a lot of responses, of course:

locate clang

/data/sauvegardes/dev/Java/Experimentations/Angular4/bikes/node_modules/blocking-proxy/.clang-format
/data/sauvegardes/dev/Java/Experimentations/Angular4/bikes/node_modules/node-gyp/gyp/tools/Xcode/Specifications/gyp.xclangspec
/data/sauvegardes/dev/Java/Experimentations/Angular6/ng6-proj/node_modules/blocking-proxy/.clang-format
/data/sauvegardes/dev/Java/Experimentations/Angular6/ng6-proj/node_modules/node-gyp/gyp/tools/Xcode/Specifications/gyp.xclangspec
/data/sauvegardes/dev/Java/Experimentations/blog-demo/node/node_modules/npm/node_modules/node-gyp/gyp/tools/Xcode/Specifications/gyp.xclangspec
/data/sauvegardes/dev/Java/Experimentations/blog-demo/node_modules/blocking-proxy/.clang-format
/data/sauvegardes/dev/Java/Experimentations/blog-demo/node_modules/node-gyp/gyp/tools/Xcode/Specifications/gyp.xclangspec
/data/sauvegardes/dev/Java/Experimentations/ol-ext-angular/.metadata/.plugins/ts.eclipse.ide.server.nodejs.embed.win32.win32.x86_64/node-v6.9.4-win-x64/node_modules/npm/node_modules/node-gyp/gyp/tools/Xcode/Specifications/gyp.xclangspec

(201 responses)

I piped this command, with dirname, sort and uniq to list only directories having such word in their name or carrying one or more file having it.

locate clang | xargs -L1 dirname | sort | uniq

it works...

/home/lebihan/dev/Java/comptes-france/metier-et-gestion/AdapterInboundWebEtude/etude/node_modules/node-gyp/gyp/tools/Xcode/Specifications
/home/lebihan/dev/Java/comptes-france/metier-et-gestion/AdapterInboundWebEtude/etude/node/node_modules/npm/node_modules/node-gyp/gyp/tools/Xcode/Specifications
/usr/include/boost/align/detail
/usr/include/boost/config/compiler
/usr/include/boost/predef/compiler
/usr/lib/linux-kbuild-6.1/scripts
/usr/lib/llvm-14/lib
/usr/lib/postgresql/14/lib/bitcode/postgres/commands
/usr/lib/x86_64-linux-gnu
/usr/local/go/misc/ios
/usr/local/go/src/debug/dwarf/testdata
/usr/local/go/src/debug/elf/testdata
/usr/local/go/src/debug/macho/testdata
/usr/share/doc
/usr/share/doc/libclang1-14
/usr/share/doc/libclang-cpp14

(108 responses)

But does _Linux_ have a command doing the same, more easily?

I have seen several references to the slocate command [\[1\]][1] [\[2\]][2] as a more secure alternative to the locate command. However I cannot install it using sudo apt install slocate. I also cannot find any references on how to install it for Ubuntu and only one reference to installing it on any Linux platform at all . But here is a question about how to use it and it is also referenced in official Ubuntu manual . So, why is it referenced, but not available? Must it be installed from source only ?

> plocate is a locate(1) based on posting lists, giving much faster searches on a much smaller index. It is a drop-in replacement for mlocate in nearly all aspects, and is fast on SSDs and non-SSDs alike. But I'm having problem using it ***under WSL1***.

$ plocate java
/var/lib/plocate/plocate.db: No such file or directory

$ sudo updatedb
/var/lib/plocate/: Is a directory

$ updatedb.plocate
/var/lib/plocate/: Is a directory

$ apt-cache policy plocate
plocate:
  Installed: 1.1.8-2+deb11u1
  Candidate: 1.1.8-2+deb11u1
  Version table:
     1.1.13-1~bpo11+1 100
        100 http://ftp.debian.org/debian  bullseye-backports/main amd64 Packages
 *** 1.1.8-2+deb11u1 500
        500 http://deb.debian.org/debian  bullseye/main amd64 Packages
        100 /var/lib/dpkg/status

$ lsb_release -a
No LSB modules are available.
Distributor ID: Debian
Description:    Debian GNU/Linux 11 (bullseye)
Release:        11
Codename:       bullseye

How is it possible for updatedb to run very fast by avoiding doing full disk scans? Since mtime of parent folder is not affected by files in subfolders. In Why does the updatedb program run so fast? > @hugomg I don't think it actually does. It should only change the > mtime of its immediate parent. – @Kusalananda I searched but cannot find the answer.