Built in Variables

In this guide, you’ll learn how to leverage awk’s built-in variables to parse and manipulate text data efficiently. We’ll cover:

• Positional variables ($1, $2, …)
• NR (Number of Records)
• NF (Number of Fields)
• $NF (Last Field in the Current Record)
• FILENAME

Throughout this article, we’ll use a sample file, size.txt, generated from df -h:

df -h > size.txt

1. Positional Variables

By default, awk splits each input line on whitespace. You can print specific fields using $1, $2, $3, and so on:

awk '{ print $1, $2, $3 }' abc.txt

• $1 → first column
• $2 → second column
• $3 → third column

For example, given abc.txt:

abc def ghi
jkl mno pqr
stu vwx yz

awk '{ print $1 }' abc.txt
# Output:
abc
jkl
stu

awk '{ print $2 }' abc.txt
# Output:
def
mno
vwx

awk '{ print $3 }' abc.txt
# Output:
ghi
pqr
yz

2. NR: Number of Records

NR tracks the current record (line) number. This is useful for adding line numbers or filtering specific lines:

awk '{ print NR, $1, $4 }' size.txt

Produces:

1 Filesystem Avail
2 udev       1.8G
3 tmpfs      356M
4 /dev/sda1  9.3G
...

To print only the 8th line’s available space:

awk 'NR == 8 { print $4 }' size.txt
# Output:
360G

3. NF: Number of Fields

NF contains how many fields are in the current record. It helps you spot inconsistencies in your data:

awk '{ print NF }' size.txt

7
6
6
6
6
6
6
6
6

4. Combining NR and NF

Print both the record number and its field count:

awk '{ print NR, NF }' size.txt

1 7
2 6
3 6
4 6
...

Concatenate text with values:

awk '{ print "Line", NR, "has", NF, "fields" }' size.txt

Line 1 has 7 fields
Line 2 has 6 fields
...

5. $NF: The Last Field

Use $NF to refer directly to the last field of each record:

awk '{ print $NF }' size.txt

on
/dev
/run
/
...

6. FILENAME: Current Filename

FILENAME holds the name of the file being processed (empty when reading from stdin):

awk '{ print FILENAME, $1 }' size.txt

size.txt Filesystem
size.txt udev
size.txt tmpfs
...

When you pipe data into awk, FILENAME is empty:

df -h | awk '{ print FILENAME, $1 }'
# Output:
 Filesystem
 udev
 tmpfs
 ...

7. Summary Table of Built-in Variables

8. Custom Field Separators

If your data uses a delimiter other than whitespace, set the -F option:

awk -F, '{ print $1, $NF }' data.csv

Links and References

• GNU awk Manual
• awk User’s Guide
• Korn Shell: awk Built-in Variables