grep with context

danielbmartin · 04-23-2015, 06:54 AM

A sample input file:

Code:

Audi
Bentley
Chevrolet
Dodge
QQQ
Ford
Honda
Isuzu
Jaguar
QQQ
Kaiser
Lexus

This InFile contains many lines, and a few contain QQQ.

Case 1: To extract only those lines with QQQ

Code:

grep "QQQ" $InFile

Case 2: The negative of Case 1. To extract all lines which do not contain QQQ

Code:

grep -v "QQQ" $InFile

Case 3: To extract the lines with QQQ and one preceding line

Code:

grep -B1 "QQQ" $InFile

Cases 1, 2, and 3 work as expected. Now, an attempt to extract the negative of Case 3

Code:

grep -v -B1 "QQQ" $InFile

This does not work as expected. It returns the entire input file.

Please explain.

Daniel B. Martin

SoftSprocket · 04-23-2015, 08:29 AM

Have you thought about it? You are telling grep to find every line but one and telling it to return those lines and the lines before them as well.

TobiSGD · 04-23-2015, 08:52 AM

Quote:

Originally Posted by SoftSprocket

Have you thought about it? You are telling grep to find every line but one and telling it to return those lines and the lines before them as well.

When you think about it, it should return every line but the last one and the ones containing QQQ+the line before twice. Problem is, it doesn't, and I couldn't find any explanation for that either.

millgates · 04-23-2015, 09:01 AM

from info grep:

Code:

2.1.5 Context Line Control
--------------------------

Regardless of how these options are set, `grep' will never print any
given line more than once.  If the `-o' (`--only-matching') option is
specified, these options have no effect and a warning is given upon
their use.

schneidz · 04-23-2015, 09:01 AM

is there a bugzilla that can be logged somewhere. it probably is functions as designed but it would be nice to have the developers explanation.

TobiSGD · 04-23-2015, 09:50 AM

Quote:

Originally Posted by millgates

from info grep:

Code:

2.1.5 Context Line Control
--------------------------

Regardless of how these options are set, `grep' will never print any
given line more than once.  If the `-o' (`--only-matching') option is
specified, these options have no effect and a warning is given upon
their use.

Thanks. So the problem is not a software problem, but a documentation problem, this information is simply not present on the manpage (and to be quite frank, I never use info, I find it uncomfortable and hard to use).

millgates · 04-23-2015, 10:03 AM

I don't like the info pages myself and I only use them as a last resort when I can't find what I'm looking for in the man pages, just for case it happens to contain that one extra sentence that explains the problem.

TobiSGD · 04-23-2015, 10:13 AM

Quote:

Originally Posted by millgates

I don't like the info pages myself and I only use them as a last resort when I can't find what I'm looking for in the man pages, just for case it happens to contain that one extra sentence that explains the problem.

I will keep that in mind. It never came to my mind that the manpages might not contain the same information as the info pages do.

millgates · 04-23-2015, 10:39 AM

They usually do, but not always. Often, the man page ends with a reference to info. For example, all the way down my grep manpage says:

Code:

   TeXinfo Documentation
       The  full  documentation  for  grep  is maintained as a TeXinfo manual,
       which you can read at http://www.gnu.org/software/grep/manual/.  If the
       info and grep programs are properly installed at your site, the command

              info grep

       should give you access to the complete manual.

NOTES
       This  man  page  is maintained only fitfully; the full documentation is
       often more up-to-date.

To get back to the original question, I would suggest an alternative to the non working grep:

Code:

awk -F'[^\n]*\nQQQ\n' -vORS= -vOFS= -vRS='\f' '($1=$1)||1' <inFile

makyo · 04-25-2015, 07:49 AM

Hi.

I often use cgrep, and it appears to work as one might expect:

Code:

#!/usr/bin/env bash

# @(#) s1	Demonstrate "inverse" pattern match, cgrep.
# cgrep, see:
# http://sourceforge.net/projects/cgrep/

# Utility functions: print-as-echo, print-line-with-visual-space, debug.
# export PATH="/usr/local/bin:/usr/bin:/bin"
LC_ALL=C ; LANG=C ; export LC_ALL LANG
pe() { for _i;do printf "%s" "$_i";done; printf "\n"; }
pl() { pe;pe "-----" ;pe "$*"; }
db() { ( printf " db, ";for _i;do printf "%s" "$_i";done;printf "\n" ) >&2 ; }
db() { : ; }
C=$HOME/bin/context && [ -f $C ] && $C cgrep

FILE=${1-data1}

pl " Input data file $FILE:"
cat $FILE

pl " Results, cgrep:"
cgrep -D -1 QQQ $FILE

pl " Results, inverse cgrep:"
cgrep -D -V -1 QQQ $FILE

pl " Results, inverse cgrep, wrong -v (lowercase), expect failure:"
cgrep -D -v -1 QQQ $FILE

exit 0

producing:

Code:

$ ./s1

Environment: LC_ALL = C, LANG = C
(Versions displayed with local utility "version")
OS, ker|rel, machine: Linux, 2.6.26-2-amd64, x86_64
Distribution        : Debian 5.0.8 (lenny, workstation) 
bash GNU bash 3.2.39
cgrep ATT cgrep 8.15

-----
 Input data file data1:
Audi
Bentley
Chevrolet
Dodge
QQQ
Ford
Honda
Isuzu
Jaguar
QQQ
Kaiser
Lexus

-----
 Results, cgrep:
Dodge
QQQ
Jaguar
QQQ

-----
 Results, inverse cgrep:
Audi
Bentley
Chevrolet
Ford
Honda
Isuzu
Kaiser
Lexus

-----
 Results, inverse cgrep, wrong -v (lowercase), expect failure:
Audi
Bentley
Chevrolet
Dodge
QQQ
Ford
Honda
Isuzu
Jaguar
QQQ
Kaiser
Lexus

One needs to compile it. I have done that in both 32-bit and 64-bit without problems.

Best wishes ... cheers, makyo