I want to check the number of bytes on each line of a file.
If the number of bytes is not the expected number, then store the record along with its line number in a new file.
Asked
Active
Viewed 166 times
1 Answers
1
Try:
expected=72
<a-file LC_ALL=C grep -nxvE ".{$expected}" >a-new-file
Or:
<a-file LC_ALL=C awk -v expected=72 '
length != expected {print FNR, $0}' >a-new-file
length returns the length without the line delimiter. Beware some grep/awk implementations would choke on input that contains NUL bytes or very long lines.
GNU awk won't and can get the length of the record including the delimiter with:
<a-file LC_ALL=C gawk -v expected=73 '
length($0 RT) != expected {print FNR, $0}' >a-new-file
Or you could use perl:
<a-file perl -nse '
print "$. $_" unless length($_) == $expected' -- -expected=73 >a-new-file
Add the -l option if you don't want to take into account the line delimiter. perl works with bytes by default regardless of the locale, so you don't need the LC_ALL=C.
Stéphane Chazelas
- 544,893
-
Need it using Unix, for loop, if else and grep to be more specific. – Anaya Sep 30 '22 at 07:31
-
@Anaya, certainly not. Using shell loops to process text is very bad practice. – Stéphane Chazelas Sep 30 '22 at 07:34
-
@Anaya please edit your question and explain exactly what you need. Doing this in the shell is a bad idea. – terdon Sep 30 '22 at 07:59