find
Last updated: 5 Aug 2020 ·
Posted in wiki#notes
find
Find by name
find -name "query"
find -iname "query" # Ignore case
find -not -name "query_to_avoid"
find \! -name "query_to_avoid"
Find by type
find -type <type_descriptor> <query> # Basic usage
# Common type descriptors:
# f: File
# d: Directory
# l: Symlink
# c: Character device
# b: Block device
find /usr -type f -name "*.conf" # Search for files ending in `.conf` in `/usr` directory
find /usr -type f -and -name "*.conf" # Equivalent to above; `-and` combines two queries
find -name query_1 -or -name query_2 # `-or` returns results that match either expression
Find by size
find -size <number><suffix> # Basic usage
# Common suffixes
# c: Bytes
# k: Kilobytes
# M: Megabytes
# G: Gigabytes
# b: 512-byte blocks
find /usr -size 50c # Find all files in `/usr` exactly 50 bytes in size
find /usr -size -50c # Less than 50 bytes in size
find /usr -size +600M # More than 600 megabytes in size
Find by time
- Access time (
atime): The last time a file was read or written to. - Modification time (
mtime): The last time the contents of the file were modified. - Change time (
ctime): The last time the file’s inode metadata was changed.
find /usr -mtime 1 # Files modified within the last day
find /usr -atime -1 # Accessed less than a day ago
find /usr -ctime +3 # Meta information changed more than 3 days ago
find /usr -mmin -1 # Modified within the last minute
find / -newer reference_file # Created/changed more recently than reference_file
Find by owner
find / -user <user>
find / -group <group>
Find by permissions
find / -perm 644
find / -perm -644 # Files with at least these permisions
Find by depth
find -mindepth <num>
find -maxdepth <num>
Execute commands on find results
Other commands can be executed on results returned by find using the -exec/-delete options or by piping the output to the xargs command or GNU Parallel.
find <find_parameters> -exec <command_and_options> {} \;
find . -type d -perm 755 -exec chmod 700 {} \; # Change directory permissions from 755 to 700
find . -name "*.json" -delete # Delete files ending in `.json`
# Useful options
# -exec command {} +: Build command by appending each selected file name at the end, and then execute it
# -execdir: Run command from the subdirectory containing the matched file
find ./docs -type f -print | xargs rm
# Find all files in `./docs` and remove them
find . -name "*.foo" -print0 | xargs -0 grep bar
# Use the null character to delimit file names (necessary for dealing with filenames with `,` or space)
find . -name "*.foo" -print0 | parallel -0 grep bar
# Equivalent to above
locate
Note: locate is faster because it searches through a prebuilt database of files generated by the updatedb command or a cronjob automatically created during installation that runs every 24 hours. To manually update the database, use sudo updatedb.
locate <query> # Return files whose full path names contain the query
locate -b <query> # Return only files whose names contain the query
locate -n 10 *.py # Limit to 10 results; Wildcard search
locate -i readme.md # Case-insensitive search
locate -c *.md # Display number of found entries
locate -S # Print statistics about each used database
whereis
whereis command can be used to efficiently locate the binary, source, and manual page files for a command.
> whereis rg
rg: /usr/bin/rg /usr/share/man/man1/rg.1.gz
which
which searches for the binary for a command in your PATH.
> which rg
/usr/bin/rg
fd
Info: oh-my-zsh's common-aliases plugin aliases fd to find . -type d -name. To avoid the clash between this alias and fd, put unalias fd or alias fd='\fd' at the end of your .zshrc.
Basic usage
fd [options] <pattern> <path> # Basic usage
fd # List all files in current directory recursively, similar to `ls -R`
fd netfl # Basic search; Recursively search current directory for the pattern `netfl`
fd '^x.*rc$' # Regex search; Search for entries that start with `x` and end with `rc`
fd '^x.*rc$' /etc # Search in `/etc` directory
fd -g libc.so /usr # Glob-based search; Find all `libc.so` files in `usr`
# Useful options
# -H, --hidden: Search in hidden directories
# -I, --no-ignore: Search directories and show files that match `.gitignore` patterns
# -p, --full-path: Search in full paths instead of just filenames
# -l, --list-details: Use a long listing format with file metadata
# -L, --follow: Follow symbolic links
Find by type
fd -t <filetype> <pattern>
# Common filetypes:
# f: File
# d: Directory
# l: Symlink
# x: Executable
# e: Empty
# s: Socket
# p: Pipe
Find by size
fd -s <size> <pattern>
Find by time
fd --changed-within <date|dur> <pattern> # Filter by file modification time (newer than)
fd --changed-before <date|dur> <pattern> # Filter by file modification time (older than)
Find by depth
fd -d <depth> <pattern> # Set maximum search depth
Find by extension
fd -e json # Find files with `.json` extension
fd -e json <pattern> # Find json files that contain the pattern
Exclude files and directories
fd -E '*.zip' <pattern> # Exclude zip files
fd -H -E .git <pattern> # Search in hidden dirs but exclude matches from `.git`
# You can create a `.fdignore` file (similar to `.gitignore`) to make exclude-patterns permanent.
Execute commands on fd results
# -x, --exec <cmd>: Execute a command for each search result
# -X, --exec-batch <cmd>: Execute a command with all search results at once
fd -e zip -x unzip # Recursively find and unzip all zip archives
fd -e h -e cpp -x clang-format -i # Find all `*.h` and `*.cpp` files and auto-format them inplace with `clang-format -i`
fd -g 'test_*.py' -X vim # Find all `test_*.py` files and open them with `vim`
# https://vimhelp.org/usr_07.txt.html#07.2
fd -H '^\.DS_Store$' -tf -X rm # Recursively remove all `.DS_Store` files
Placeholder tokens
fd -e jpg -x convert {} {.}.png # Convert all `*.jpg` files to `*.png` files
# `{}` will be replaced by the path of the search result (`docs/images/party.jpg`)
# Other placeholder tokens
# {.}: Like {}, but without the file extension (`docs/images/party`)
# {/}: The basename of the search result (`party.jpg`)
# {//}: The parent of the discovered path (`docs/images`)
# {/.}: The basename without the extension (`party`)