CODEWITHSUNDEEP
bashcsvawkfield
Rusty Lemur
Rusty Lemur

Reputation: 1875

How to call awk from bash with last field as variable

Suppose I have a CSV input file called input.csv:

1,2,3,valueIWantToPrint
3,5,2,valueIWantToPrint

I currently need to print the last element of each row of that input with awk, which is easily accomplished with the field separator and NF variables:

awk -F"," '{print $NF}' input.csv

But now let's say that I want to make the field that I want to print a variable, because later perhaps the input format will change and it will be a different field.

Input file:

1,2,valueIWantToPrint,3
3,5,valueIWantToPrint,2

Script:

FIELD_TO_PRINT=3
awk -F"," -v fieldToPrint=FIELD_TO_PRINT '{print $fieldToPrint}' input.csv

Ok, that was easy. But now to make it as flexible as possible, I would like have the ability to set FIELD_TO_PRINT to the equivalent of NF so that I can print the last value regardless of the number of fields. What I'm after is something like this:

Input:

1,2,3,7,2,5,23,1,3,6,valueIWantToPrint
3,5,2,6,3,valueIWantToPrint

This script doesn't work, but illustrates what I am trying to accomplish:

FIELD_TO_PRINT=NF
awk -F"," -v fieldToPrint=FIELD_TO_PRINT '{print $fieldToPrint}' input.csv

Is there a convenient way to set a variable to mean "the last field in record?" This example is pretty trivial, but ultimately the FIELD_TO_PRINT variable will be put in a separate configuration file, and the awk script will be much larger and more complex. So having a good way to accomplish this will be very useful.

Upvotes: 3

Views: 171

Answers (2)

karakfa
karakfa

Reputation: 67507

you can use this trick

$ awk -F, -v n=-1 '{print (n<0)?$(NF+n+1):$n}' file

valueIWantToPrint
valueIWantToPrint

assume negative indices start counting from NF backwards, so -2 will mean the penultimate field etc.

Upvotes: 3

anubhava
anubhava

Reputation: 785156

You can use this round-about way:

Negative n means NF:

n=-1
awk -F, -v n="$n" '{print (n<0 ? $NF : $n)}' f.csv


valueIWantToPrint
valueIWantToPrint

and when n > 0 print numbered field:

n=3
awk -F, -v n="$n" '{print (n<0 ? $NF : $n)}' f.csv


3
2

Upvotes: 3

Related Questions