Print a column from a CSV with awk (and when not to)

awk -F, '{print $2}' file.csv pulls out a column in one line, and for machine-generated CSV that's the end of the story. For CSV that came from a spreadsheet, there's a trap that produces wrong numbers instead of errors — so it's worth knowing where the line is.

1. One column

$ clai print the first column of data.csv→ awk -F, '{print $1}' data.csvnameAliceBob

-F, sets the field separator. Fields are $1, $2, … and $0 is the whole line.

2. Several columns, skipping the header

$ clai print columns 1 and 3 from data.csv without the header row→ awk -F, 'NR>1 {print $1, $3}' data.csvAlice 100Bob 250

NR>1 skips line one. NR is the record number, and using it as a condition is how awk does "for the rows where…".

3. Keep it comma-separated on the way out

$ clai extract columns 1 and 3 as valid csv→ awk -F, 'BEGIN{OFS=","} NR>1 {print $1,$3}' data.csvAlice,100Bob,250

print $1, $3 joins with OFS, which defaults to a space. Set it in BEGIN and the output is CSV again.

4. Sum a column

$ clai total up the third column of data.csv→ awk -F, 'NR>1 {s+=$3} END {print "total", s}' data.csvtotal 350

No variable declaration, no initialisation — awk starts numbers at zero. END runs once after the last line.

5. Filter rows, then take a column

$ clai show the names where the amount is over 200→ awk -F, 'NR>1 && $3 > 200 {print $1}' data.csvBob

This is where awk beats a spreadsheet: the same one-liner runs over a 3 GB file without opening it.

6. The trap: a comma inside a quoted field

$ clai print the second column→ awk -F, 'NR>1 {print $2}' data.csvBerlin"Paris

The row Bob,"Paris, France",250 has three logical fields but four commas. Awk doesn't know about quoting, so $2 becomes "Paris, $3 becomes France", and the sum from step 4 silently returns 100 instead of 350. It doesn't error — it just gives you the wrong number.

Gotchas

• Check for quotes before you trust a total. grep -c '"' file.csv — if that's zero, plain awk is safe. If it isn't, use a real CSV parser: csvcut -c 1,3 data.csv (csvkit), mlr --csv cut -f name,amount (Miller), or three lines of Python.
• -F, is not -F', '. A separator of comma-plus-space silently produces empty fields on rows without the space.
• Windows line endings leave \r on the last field. $NF compares unequal for no visible reason. Fix with awk -F, '{gsub(/\r$/,"")} …' or run dos2unix first.

Related questions

How do I print the last column? awk -F, '{print $NF}'. NF is the number of fields, so $NF is the last and $(NF-1) the one before it.

How do I use a tab separator? awk -F'\t' — quote it so the shell doesn't eat the backslash.

Is awk faster than pandas for this? For extracting or summing a column from a large file, yes, by a lot — it streams and never builds a data structure. For anything involving joins or reshaping, use pandas.

See also

• Find and replace across every file in a project
• Renaming hundreds of files: from for loops to plain English
• From one-liner to checked-in script: when to graduate

Describe the extraction you want; CliAI writes the awk program, including the BEGIN{OFS=","} you'd forget. Install it in one line.