What awk does and why you'd use it for tables
Awk is a text processing tool that reads lines from a file or command output, splits each line into columns, and lets you rearrange, filter, or format those columns. When your data arrives as space-separated or comma-separated values, awk can turn it into a properly aligned table with headers, padding, and consistent column widths — all without opening a spreadsheet.
The tool is built into every Linux and Unix system, and it works on macOS and Windows (via WSL or Git Bash). You run it from the command line, which means you can pipe data directly from network tools, log files, or other commands into a formatted table in seconds.
A typical use case: you run a command like netstat or ip addr to see network interfaces or connections, but the output is hard to read because columns are misaligned. Awk can parse that output and rebuild it as a clean, readable table with even spacing.
Key Takeaways
- Awk reads input line by line, splits each line into fields (columns), and lets you print, filter, or rearrange those fields in any order.
- The basic syntax is awk '{print $1, $2, $3}' to print specific columns, or awk -F: '{print $1}' to change the field separator from space to colon.
- Use printf instead of print to control column width and alignment, which is how you create a properly formatted table.
- You can add headers, filter rows with conditions, and calculate totals or counts — all in a single awk command without writing a script file.
- Awk works on any text data: command output, CSV files, log files, or system information — making it useful for turning messy network or system data into readable tables.
The basic syntax: fields, separators, and printing columns
Awk treats each line of input as a record and splits it into fields (columns) based on a separator. By default, the separator is whitespace — any number of spaces or tabs. Each field is numbered: $1 is the first column, $2 is the second, and so on. $0 is the entire line.
The simplest awk command prints specific columns. If you have a file with three columns and want to print only the first and third:
awk '{print $1, $3}' filename.txt
If your data uses a different separator — like a colon in /etc/passwd or a comma in a CSV file — use the -F flag to change it:
awk -F: '{print $1, $3}' /etc/passwd
This reads /etc/passwd, splits each line on colons, and prints the first and third fields (username and user ID). The comma between $1 and $3 in the print statement adds a space between them in the output.
Creating aligned columns with printf
The print statement adds a single space between fields, which looks messy if columns have different widths. To create a proper table with aligned columns, use printf instead. Printf lets you specify the exact width and alignment of each column.
The syntax is printf "format string", field1, field2, field3. In the format string, %-15s means "print this field left-aligned in a space 15 characters wide". %10d means "print this field right-aligned as a number in 10 characters". Here's an example:
awk '{printf "%-15s %10s %8s\n", $1, $2, $3}' data.txt
This prints three columns: the first left-aligned in 15 characters, the second right-aligned in 10 characters, and the third right-aligned in 8 characters. The \n at the end adds a newline after each row. If you run this on a file with misaligned data, the output will be a clean, readable table.
Adding headers and formatting the output
A table is easier to read with a header row. You can add one using the BEGIN block, which runs before awk processes any input:
awk 'BEGIN {printf "%-15s %10s %8s\n", "Name", "Value", "Count"} {printf "%-15s %10s %8s\n", $1, $2, $3}' data.txt
The BEGIN block prints the header row once, then the main block prints each data row with the same formatting. If you want a line of dashes under the header, add it in the BEGIN block:
awk 'BEGIN {printf "%-15s %10s %8s\n", "Name", "Value", "Count"; printf "%s\n", "-------------------------------------------"} {printf "%-15s %10s %8s\n", $1, $2, $3}' data.txt
The semicolon separates statements within the BEGIN block. You can make the dash line as long as you need — it should be at least as wide as your table.
Filtering rows and working with command output
Awk can filter rows based on conditions. For example, to print only rows where the second column is greater than 100:
awk '$2 > 100 {printf "%-15s %10s\n", $1, $2}' data.txt
The condition $2 > 100 comes before the action block. Only rows that match are printed. You can also filter by text: awk '$1 == "eth0"' prints only rows where the first field equals "eth0".
To format output from a command like netstat or ip addr, pipe it directly into awk:
netstat -an | awk 'NR > 2 {printf "%-20s %15s %15s\n", $4, $5, $6}'
The NR > 2 condition skips the first two lines (which are usually headers or blank). NR is the row number, so this starts printing from row 3. The pipe | sends the output of netstat directly to awk, so you don't need a file.
Calculating totals and counts within awk
Awk can sum columns, count rows, or calculate averages — all while formatting the output. To sum the second column and print the total at the end, use an END block:
awk '{sum += $2; printf "%-15s %10d\n", $1, $2} END {printf "%-15s %10d\n", "Total", sum}' data.txt
The sum += $2 adds each row's second field to a running total. The END block runs after all input is processed and prints the total. You can do the same with counts: count++ increments a counter for each row, and END {print "Rows:", count} prints the final count.
This approach is useful for network logs or system data where you want to see individual entries formatted as a table, plus a summary line at the bottom showing totals or averages.
Common awk patterns for network and system data
Network and system commands often produce output that's hard to parse. Here are patterns that work well with awk. For ip addr output, which shows network interfaces and IP addresses:
ip addr | awk '/inet / {printf "%-15s %s\n", $NF, $2}'
The /inet / condition matches only lines containing "inet", and $NF is the last field (the interface name). For ps aux output, which shows running processes:
ps aux | awk 'NR > 1 {printf "%-10s %8s %8s %s\n", $1, $3, $4, $11}'
This skips the header row and prints username, CPU percentage, memory percentage, and the command name. For log files with timestamps, you can extract and reformat the date:
awk '{printf "%s %s %s\n", $1, $2, $3}' logfile.txt
These patterns work because awk handles the field splitting automatically — you just specify which fields to print and how to format them.
Frequently Asked Questions
What's the difference between print and printf in awk?
Print adds a space between fields and a newline at the end, but doesn't let you control spacing or alignment. Printf uses a format string where you specify exact column widths and alignment, so you get a properly formatted table. Use print for quick output and printf when you need readable columns.
How do I change the field separator if my data uses commas or colons?
Use the -F flag: awk -F, '{print $1, $2}' for comma-separated data, or awk -F: '{print $1, $3}' for colon-separated data. You can also set it inside the script: awk 'BEGIN {FS=","} {print $1, $2}'. The FS variable controls the field separator.
Can I save an awk command to a file and run it multiple times?
Yes. Save your awk code to a file (for example, format.awk) and run it with awk -f format.awk filename.txt. This is useful if your formatting command is long or you use it regularly. The file should contain just the awk code, without quotes.
How do I skip the header row when processing command output?
Use the NR > 1 condition to skip the first row, or NR > 2 to skip the first two. NR is the row number. For example: netstat -an | awk 'NR > 2 {print $1, $2}' skips the first two lines of netstat output and prints columns 1 and 2 from the rest.
What if a field contains spaces or special characters?
If a field itself contains spaces, the default whitespace separator will split it into multiple fields. Use -F to specify a different separator that doesn't appear in your data, or use a regex pattern like -F'[ ]+' to split on multiple spaces only. For CSV files with quoted fields, awk becomes harder to use — consider cut or a scripting language instead.