awk/gawk seems to overwrite columns on result

Question

[TRIVIAL] [SOLVED]

tl:dr - DOW CRLF made to feel as though awk misbehaved; trivial mistake!

Am not sure why awk/gawk is behaving this way. Any hints? Ever seen awk results overwriting the columns on terminal? I am seeing this on just one machine; everywhere else, the stuff works just fine! Not sure if it's the terminal misbehaving or any other thing. An stty sane doesn't seem to help either. Anyone seen a similar thing?

 $ cat in.txt 
2132088 AAA_WEB_50181_552222195-1
2110665 AAA_WEB_50177_547796499-2
2157591 AAA_WEB_50181_551310844-1
2117370 AAA_WEB_50178_549250104-1
2109777 AAA_WEB_50180_551512637-1
2112704 AAA_WEB_50179_410838823-13
2116984 AAA_WEB_50178_549463801-2
2119511 AAA_WEB_50179_550076677-2
2122008 AAA_WEB_50180_551679428-1
2124984 AAA_WEB_50177_548026293-1
 $ awk '{print $2 " " $1}' in.txt 
 213208850181_552222195-1
 211066550177_547796499-2
 215759150181_551310844-1
 211737050178_549250104-1
 210977750180_551512637-1
 211270450179_410838823-13
 211698450178_549463801-2
 211951150179_550076677-2
 212200850180_551679428-1
 212498450177_548026293-1
 $ cat /etc/issue.net 
Ubuntu 14.04.4 LTS
 $ gawk '{print $2 " " $1}' in.txt 
 213208850181_552222195-1
 211066550177_547796499-2
 215759150181_551310844-1
 211737050178_549250104-1
 210977750180_551512637-1
 211270450179_410838823-13
 211698450178_549463801-2
 211951150179_550076677-2
 212200850180_551679428-1
 212498450177_548026293-1
 $ gawk -W version
GNU Awk 4.0.1
Copyright (C) 1989, 1991-2012 Free Software Foundation.

Screenshot:

@triplee - yep! Figured it out just now! Thanks! – raghava Jun 01 '16 at 15:47 — raghava, Jun 01 '16 at 15:47

Inian · Accepted Answer · 2020-06-30T12:59:14.513

6

As pointed out by tripleee, the issue is likely due to DOS line terminators, a simple fix could be to strip the special characters using tr and feed it to awk for processing.

< in.txt tr -dc '[:print:]\n' |  gawk '{print $2 " " $1}'

In the above example, tr -dc '[:print:]\n' allows only the printable characters from the input file before feeding it to awk.

edited Jun 30 '20 at 12:59

answered Jun 01 '16 at 15:44

Inian

80,270
14
142
161

Exactly! Figured out this as well. Ran a dos2ux and then got it fixed. Should have seen that coming! :D – raghava Jun 01 '16 at 15:48

sjsam · Answer 2 · 2016-06-01T15:59:43.683

0

The problem could be with the carriage return (\r) present at the end of the lines, if the file is coming from a dos-like environment.

A gawk only solution would be.

gawk '{printf "%s %s\n",gensub(/\r$/,"",-1,$2),$1}' in.txt

edited Jun 01 '16 at 15:59

answered Jun 01 '16 at 15:41

sjsam

21,411
5
55
102

awk/gawk seems to overwrite columns on result

2 Answers2