LinuxQuestions.org - [SOLVED] awk power

- Programming (https://www.linuxquestions.org/questions/programming-9/)

- - awk power (https://www.linuxquestions.org/questions/programming-9/awk-power-948175/)

I have 2 files as attached.

f1 contain a lots of lines like this:

Quote:

319030003003~319030084005~1~0070~EA~
399056084020~319030084006~1~0080~EA~
319056003001~399056084019~1~0010~EA~
319030084005~399056084020~1~0020~EA~
319030084006~319056084001~1~0030~EA~
319030003089~319030084005~1~0070~EA~
399056084029~319030084006~1~0080~EA~

f2 contains

Quote:

319030084006~
319056003001~
319030084005~

So using awk I want to delete the lines in f1, in which the 1st column contains the number in f2 and save in f3.

I want in f3:

Quote:

319030003003~319030084005~1~0070~EA~
399056084020~319030084006~1~0080~EA~
319030003089~319030084005~1~0070~EA~
399056084029~319030084006~1~0080~EA~

How can I do this using awk ??

Code:

awk -F~ 'FNR == NR { _[$1]++ } FNR < NR && !_[$1]' f2 f1 > f3

This uses the TRUE and FALSE interpretation of awk: reading f2 it increases the value of the array _ with index $1; if the number has been encountered in f2, the value of _[$1] is greater or equal to 1, that is TRUE, hence the negation is FALSE and the matching record in f1 is not printed out.

On the contrary, when it encounters a line in f1 whose first field was not listed in f2, the evaluation of _[$1] returns an empty string (that is FALSE) and the negation makes it TRUE: indeed we want these lines be printed out!

Hope this helps.

Just for fun, here's a quick&dirty solution using grep, sed, and bash's process substitution.

Code:

grep -v -f <( sed 's/^/^/' f2 ) f1

The sed p.sub is there to alter the contents of f2 first into regexes that only match the beginning of the line.

Although the question was for awk:

Code:

$ join -v 1 -t "~" <(sort f1) <(sort f2)