CODEWITHSUNDEEP
shellawk
stacktome
stacktome

Reputation: 790

How to grep the word next to the search word (....) using Shell script?

I wrote a shell script to identify given word and grep the word next to the given word. Please see my sample input file below. Also I need to write the same file as shown below in my sample output file. 

#!/bin/bash 

LOGFILE=$1 

if [ -f $LOGFILE ]
then 
        sed -i.bak 's/sqrt(\(\w*\))/sqrt(\1) print(\1)/g' $LOGFILE
else 
        echo "$LOGFILE not found " 
fi

but when I run this shell script nothing happens in my file. Could you please help me?

Example input file: 

wewe **sqrt(a+B)**  eeenwerer ergv. **sqrt(2x-3)** gtgnwerer ergv. **sqrt(wea*B)** gjjj nwerer ergv. **sqrt(a2/B)** rrnwerer ergv.

Expected file output: 

wewe **sqrt(a+B)** eeenwerer ergv. ***write (a+B);*** **sqrt(2x-3)** gtgnwerer ergv. ***write (2x-3);*** ** **sqrt(wea*B)** gjjj nwerer ergv. ***write (wea*B);*** ** **sqrt(a2/B)** rrnwerer ergv.***write (a2/B);*** **

Upvotes: 3

Views: 483

Answers (4)

Armali
Armali

Reputation: 19375

This is somewhat similar to Håkon Hægland's solution, but shorter and works also with old versions of awk where match() cannot have 3 arguments:

awk -F'sqrt\\(' '
    { for (i=1; ++i<=NF; )
        $i = gensub("(.*)\\)(.*)", "sqrt(\\1)\\2 write(\\1);", 1, $i)
      print
    }           ' $LOGFILE

Upvotes: 0

Bruce Barnett
Bruce Barnett

Reputation: 948

Since you want to find all characters except the ')' after 'sqrt(' - you can use '[^)]*' to match this character set. Then you can explicitly search for the patterns sqrt(...) and sqrt(...(...)....) as two separate cases.

sed '
    s/sqrt(\([^)]*)\)/sqrt(\1) print(\1)/g
    s/sqrt(\([^)]*([^)]*)[^)]*)\)/sqrt(\1) print(\1)/g
'

This is easier to understand if you change '[^)]*' to Z

sed '
    s/sqrt(\(Z)\)/sqrt(\1) print(\1)/g
    s/sqrt(\(Z(Z)Z)\)/sqrt(\1) print(\1)/g
'

Upvotes: 0

pravin_23
pravin_23

Reputation: 126

In case the words are in a separate line you cant try this ,It will search for the word and print the line next to it.

awk '/word to search :/{getline;print;}'

Upvotes: 0

H&#229;kon H&#230;gland
H&#229;kon H&#230;gland

Reputation: 40718

You can try:

awk -f mod.awk input.txt

where input.txt is you input log file, and mod.awk is:

{
    str=str $0 ORS
}
END {
    str2=""
    prev=""
    while(match(str,/sqrt\(([^)]*)\)/,a)) {
        cur=substr(str,RSTART,RLENGTH)
        end=substr(str,RSTART+RLENGTH)
        str2=str2 substr(str,1,RSTART-1) prev cur
        str=end
        prev="*write(" a[1] ");"
    }
    print str2 str prev
}

Given input file input.txt: 

wewe sqrt(a+B)  eeenwerer ergv. sqrt(2x-3) gtgnwerer ergv. sqrt(wea*B) gjjj nwerer ergv. sqrt(a2/B) rrnwerer ergv.

Running awk -f mod.awk input.txt gives output:

wewe sqrt(a+B)  eeenwerer ergv. *write(a+B);sqrt(2x-3) gtgnwerer ergv. *write(2x-3);sqrt(wea*B) gjjj nwerer ergv. *write(wea*B);sqrt(a2/B) rrnwerer ergv.*write(a2/B);

Upvotes: 1

Related Questions