Files

Listing files

ls
 
# show hidden files and folders
ls -a
 
# show files in long format
ls -l
 
# human readable file sizes
ls -h
 
# show inode numbers
ls -i
 
# can be used when ls not available
echo *

Create directory

# crashes if the path does not exist
mkdir /path/to/folder
 
# creates path if required for folder to be created
mkdir -p /path/to/folder

Copy Files

cp

• flags
  • p: preserve
  • r: recursive
  • v: verbose
  • u: update
  • a: archive

# p: preserve metadata including last modified date
# r: recursive
# v: verbose
cp -prv <source> <destination>
 
# i: interactive mode
# ask before overwriting
cp -i file1 file2
 
# copy file1 and file2 into dir
cp file1 file2 dir
 
# copy files from inside dir1 to dir2
cp dir1/* dir2
 
# u: update mode
# When copying files from one directory to another, only copy files that either don't exist or are newer than the existing corresponding files, in the destination directory
cp -u <source> <destination>
 
# a: archive mode
# Copy the files and directories and all of their attributes, including ownerships and permissions
cp -a <source> <destination>

rsync

rsync <options> <source> <destination>

Moving and Renaming files

• flags
  • i: interactive
  • u: update
  • v: verbose

mv <source> <destination>
 
# rename
mv old_name.txt new_name.txt
 
# move into directory
mv file.txt /target/directory/

Removing files

• Tip: always use ls to make sure what files are covered by the pattern
• flags
  • i: interactive mode
  • r: recursive
  • f: force
  • v: verbose

rm <file>
 
# Ask before removing file
rm -i <file>
 
# recursively remove all files from directory
rm -r <directory>
 
# silently force remove all files recursively
# ignores non-existent files
rm -rf <directory>

Create an empty file

• touch creates an empty file, or update timestamps if file already exists

touch file.txt
 
# update only access time
touch -a file.txt
 
# update only modification time
touch -m file.txt

File metadata

• stat is used to display detailed inode metadata about a file

> stat file.txt
  File: file.txt
  Size: 6               Blocks: 8          IO Block: 4096   regular file
Device: 254,3   Inode: 262398      Links: 1
Access: (0664/-rw-rw-r--)  Uid: ( 1000/ kartoos)   Gid: ( 1000/ kartoos)
Access: 2026-07-05 23:55:09.342641340 +0530
Modify: 2026-07-08 20:25:22.832577946 +0530
Change: 2026-07-08 20:25:22.832577946 +0530
 Birth: 2026-07-05 23:55:09.338641315 +0530

• Output includes:
  • Size in bytes
  • Blocks: number of 512-byte blocks allocated; IO Block: filesystem block size in bytes
    • See Block
  • File type: Derived from File Types
  • Device ID: device the file resides on (we can see major,minor from lsblk)
  • Inode ID
  • File Permissions, UID, GID
  • Hard Links count
  • Access time (atime): last time file was read
  • Modify time (mtime): last time file content was changed
  • Change time (ctime): last time inode metadata was changed (permissions, owner, link count, or content)
    • content writes also update the inode because the inode stores mtime, file size, and block pointers - all of which may change on a write
    • ctime is a superset of mtime: every mtime change triggers a ctime change, but not vice versa (e.g. chmod updates ctime only)
  • Birth time (btime): time when the file was originally created on the filesystem

Searching for files

• locate can be used to search
  • It utilizes a database index which is updated periodically (generally daily)
    • updatedb can be used manually to update db index
  • The files to be searched maybe obsolete
  • It is fast and easy

locate bin/zip

• find can be used to search for files in a directory tree
  • perform regex and wildcard pattern search
  • can perform tests like:
    • type
    • size
    • last modified
    • permission
    • belonging to user or group
  • can include logical operators for tests like or, and, not etc.
  • can perform actions
    • print (default)
    • delete
    • user defined using
      • -exec <command> {}
      • -ok <command> {}
  • options
    • min/max depth
    • avoid directories mounted on other filesystems

# by name
find /path/to/directory -name "*.txt"
 
# by type (f=file, d=directory, l=symlink)
find /path/to/directory -type f
 
# by size (>10MB)
find /path/to/directory -size +10M
 
# modified in last 7 days
find /path/to/directory -mtime -7
 
# execute command on results
find /path/to/directory -name "*.log" -exec rm {} \;

glob pattern

• * matches any character including none
• ? matches single character
• ** matches all files and subdirectories recursively
• [characters] matches any character in the set
• [!characters] or [^characters] matches any character not in the set
• [[:class:]] matches any character in member of class
  • [:alnum:]: alphanumeric
  • [:alpha:]: alphabets
  • [:digit:]: digits
    • same as [0-9]
  • [:lower:]: lowercase letters
  • [:upper:]: uppercase letters
• It is highly discouraged to use [a-z] or [A-Z] because of locale issues as well as modern linux does not follow ASCII order of characters

Links

Create symbolic links

• aka soft links
• store reference to another file or directory
• similar to shortcut
• breaks if original file is deleted
• can use relative path or absolute path
• breaks if the symlink cannot find the relative or absolute path
  • relative path can be useful to refer files inside the project so it does not break on another machine while moving whole project
  • absolute path can be useful if the original file stays permanently at one place, the symlink itself can be moved anywhere in the system and it will still work

ln -s original.txt link.txt
ln -s orig_directory link_directory
 
> rm original.txt
> cat link.txt
cat: link.txt: No such file or directory

Create hard links

• Creates another filename pointing to the same inode
  • A hard link references a file and cannot reference a directory
    • This is to avoid circular loops
  • A hard link cannot reference a file outside its own filesystem
    • ext4 file cannot hard link to any virtual filesystems like tmpfs/devtmpfs etc.
  • A hard link cannot reference to a file outside the partition
    • because inode pool is different b/w partitions
    • exception is btrfs or zfs which uses unified inode pool
• File remains accessible even if original filename is deleted
• Works even if you move the original file or the link to a different place

ln original.txt link.txt
 
# show the same inode for original.txt and link.txt
ls -li *.txt
 
> rm original.txt
> cat link.txt
<contents>

Viewing file contents

• cat
  • show all contents of file

cat <file>
 
# writing anything on stdin will be repeated to stdout
# to stop, send EOF (Ctrl + D)
> cat
hello world
hello world

• head
  • outputs top of file
• tail
  • outputs bottom of file

# first 5 lines, default is 10
head -n 5 <file>
 
# last 5 lines, default is 10
tail -n 5 <file>
 
# keep the file open and stream new lines in real time
tail -f <file>

• more
  • only moves forward
    • modern version can move backward as well
  • loads entire file before opening
• less
  • stands for “less is more”
  • improvement over more and has more features
  • can navigate forward and backward
  • loads file in pieces and hence faster
  • man, git log etc. uses it internally

more <file>
less <file>

• Keyboard shortcuts for less
  • Inspired from vi editor shortcuts
  • Page Down / Space: move one page forward
  • Page Up / b : move one page backward
  • Down Arrow / j: Scroll one line forward
  • Up Arrow / k: Scroll one line backward
  • g: beginning
  • G: end
  • h: Help screen
  • q: Quit
  • Searching
    • /pattern: Search forward
    • ?pattern: Search backward
    • After pressing enter for searching
      • n: next match
      • N: previous match
    • -i: Toggle case sensitive/insensitive

Searching within files

grep

• Global Regular Expression Print

grep [options] regex [file...]

• flags
  • i: ignore case
  • v: invert match
  • c: print count of matches
  • l: print name of each file with match
  • L: print name of each file without match
  • h: do not print file names
  • q: silent mode
  • E: extended regex

# pipe output from other program and search
ls -al /usr/bin | grep 'zip'
 
# find "fox" in multiple files
grep 'fox' sample*.txt
 
# find "fox" in multiple files
# do not print file names
grep -h 'fox' sample*.txt