dirname

TLDR

Get directory path

$ dirname [/path/to/file.txt]

Multiple paths

$ dirname -z [file1] [file2]

dirname extracts the directory portion of a pathname, removing the last component. It's commonly used in shell scripts to find the directory containing a file or script.
The command is the complement to basename.

PARAMETERS

-z, --zero

Separate output with NUL instead of newline

BEHAVIOR

Given `/path/to/file.txt`:
- dirname returns `/path/to`
Given `/path/to/`:
- dirname returns `/path`
Given `file.txt`:
- dirname returns `.`
Given `/`:
- dirname returns `/`

WORKFLOW

$ # Get directory
dirname /usr/local/bin/command
# Output: /usr/local/bin

# Get parent of current script
SCRIPT_DIR=$(dirname "$0")

# Get absolute path to script directory
SCRIPT_DIR=$(cd "$(dirname "$0")" && pwd)

# Multiple files
dirname /path/to/file1 /path/to/file2

# In scripts
cd "$(dirname "$0")" # Change to script's directory

COMMON USES

Script directory:

$ #!/bin/bash
SCRIPT_DIR=$(dirname "$0")
source "$SCRIPT_DIR/config.sh"

Parent directory:

$ parent=$(dirname "/path/to/file")

Multiple levels:

$ # Go up two directories
grandparent=$(dirname "$(dirname "/path/to/file")")

CAVEATS

Doesn't resolve symbolic links (use readlink). Doesn't check if path exists. Trailing slashes affect output. Doesn't handle all edge cases (empty strings, etc.). Path doesn't need to exist.

HISTORY

dirname has been part of Unix since the early days, included in POSIX standards alongside basename.