mjhjmtu

input:

123456.00|aswani|india|ap
23456.00|rani|us|tel
233|ramu|londan|vih

output:

aswani|ap
rani|tel
ramu|vih

I'm trying for this output with the command

awk '";OFS="print $2,$4'

but I'm getting output as:

|
rani|tel
ramu|vih

command was executing from 2 line but not from first line. i'm trying for output along with first line

You should put the first block (setting FS and OFS) inside a "BEGIN" block:

awk 'BEGIN ";OFS=" print $2,$4'

A "BEGIN" block will run before the first line of text is processed. In contrast, the stand-alone block you had will run on every line of text, but for the first line, setting FS will happen too late, since the fields will have been split already.

An alternative is to set FS through an awk command line option, but you still need to take care of OFS:

awk -F '|' '"; print $2,$4'

Or:

awk -F '|' -v OFS='|' 'print $2,$4'

UPDATE: As pointed out by @Kusalananda, some awk implementations (such as BSD awk) will re-split the fields if FS gets reset, so that setting doesn't need to happen in a "BEGIN" block. In GNU awk (which is typically the one found in Linux distributions), that does not happen, so FS needs to be set before the line is split.

As filbranden points out GNU awk (and mawk) requires FS to be correctly set before reading the first line, otherwise it will split the line on the default field separator (a sequence of whitespaces). Your code sets FS after reading each line, so it will have the wrong value while processing the first line of the file.

OpenBSD awk (at least) behaves differently and appears to split the current record when a field is accessed, but not before that. This means that your code actually works on an OpenBSD system.

I'm just going to add to that and say that to extract a set of columns from a file with no further processing, the cut command is also useful:

$ cut -d '|' -f 2,4 <file
aswani|ap
rani|tel
ramu|vih