awk change once per file

问题

awk -i inplace '
  BEGIN {FS=" "}
  BEGINFILE {changed=0} 
  { print;if ($1 == "namespace" && !changed) {print "foo";changed=1} }
' *

Is there a more elegant way to do this? Some built-in construct I missed? I am running GNU Awk 4.1.0 (and I am superb glad for -i inplace).

回答1:

awk -i inplace '$1=="namespace" && !seen[ARGIND]++ {$0=$0 ORS "foo"} 1' *

FS=" " is the default, no need to specify it explicitly.

回答2:

You could say:

... -F' ' '$1 == "namespace" && !a {$0=$0 RS "foo";a=1}1' file

or even:

... -F' ' '$1 == "namespace" && !_ {$0=$0 RS "foo";_=1}1' file

As mentioned in the comment above, this is pretty similar to what you've written except that it uses the RS variable to insert the text.

---

In order to do this for multiple files, you'd need to reset the variable:

... -F' ' 'BEGINFILE {a=0} $1 == "namespace" && !a {$0=$0 RS "foo";a=1}1' *

or

... -F' ' 'FNR==1 {a=0} $1 == "namespace" && !a {$0=$0 RS "foo";a=1}1' *

来源：https://stackoverflow.com/questions/20185553/awk-change-once-per-file