AWK to find desire lines and merge them

Question

Found easiest solution sofar to extract 3rd and 5th line after pattern (searching huge error logs)
but output I got is 2 lines and I want them merged.
Please advice.

Sample:

pattern
1st line  
2nd line  
3rd line  
4th line  
5th line
pattern 2
1st line after second pattern
2nd line after second pattern 
3rd line after second pattern 
4th line after second pattern 
5th line after second pattern

using :

awk '/pattern/ {nr[NR+3]; nr[NR+5]}; NR in nr'

I got

3rd line  
5th line
3rd line after second pattern 
5th line after second pattern

But how to get:

3rd line 5th line
3rd line after second pattern 5th line after second pattern

Thank you

Thank you noted. -v ORS=' ' works the way that it merged then all occurrences of 3rds and 5ths ending in single line. Im scanning massive log with thousand of occurrences and need output of TIME on 3rd line and 5th DESCRIPTION for each pattern match. — Ramon SK, Jun 07 '22 at 17:10

score 0 · Answer 1 · answered Jun 07 '22 at 19:50

0

Use paste like so:

awk '/pattern/ {nr[NR+3]; nr[NR+5]}; NR in nr' in_file | paste -d' ' - -

answered Jun 07 '22 at 19:50

Timur Shtatland

12,024
2
30
47

score 0 · Answer 2 · answered Jun 07 '22 at 19:59

0

You may use this awk:

awk '/^pattern/ {n=NR} NR == n+3 {printf "%s ", $0} NR == n+5' file

3rd line 5th line
3rd line after second pattern 5th line after second pattern

answered Jun 07 '22 at 19:59

anubhava

761,203
64
569
643

score 0 · Answer 3 · answered Jun 07 '22 at 20:24

0

Using sed (if an option)

$ sed -n '/^pattern/{n;n;n;h;n;n;x;G;s/\n//p}' input_file
3rd line  5th line
3rd line after second pattern 5th line after second pattern

answered Jun 07 '22 at 20:24

HatLess

10,622
5
14
32

score 0 · Accepted Answer · answered Jun 08 '22 at 12:39

Guessing at what you mean by "pattern" (see How do I find the text that matches a pattern?):

$ cat tst.awk
/^pattern/ {
    printf "%s", ors
    delete nrs
    nrs[NR+3]
    nrs[NR+5]
    ofs = ""
    ors = ORS
}
NR in nrs {
    printf "%s%s", ofs, $0
    ofs = OFS
}
END {
    printf "%s", ors
}

$ awk -f tst.awk file
3rd line 5th line
3rd line after second pattern 5th line after second pattern

RARE Kpop Manifesto · Answer 5 · 2022-06-08T14:27:54.607

0

{m,g}awk '
BEGIN { _="^"(_="[ \t]+")"|"(_)"$"
    FS = "pattern"
} $!NF = NF<=!+_ ? __ : sprintf("%s %s",
            $(getline * getline * getline * gsub(_,__)*_),        
            $(          getline * getline * gsub(_,__)*_))'

.

3rd line 5th line
3rd line after second pattern 5th line after second pattern

edited Jun 08 '22 at 14:27

answered Jun 08 '22 at 14:16

RARE Kpop Manifesto

2,453
3
11

ufopilot · Answer 6 · 2022-06-08T17:30:22.207

awk -v r1="3" -v r2="5" '
   /^pattern$/ || /^pattern 2$/{i=0; p=$0; next}
   {i++; a[p][i] = $0} 
   END{for(e in a) print a[e][r1],a[e][r2]}
' input_file

3rd line 5th line
3rd line after second pattern 5th line after second pattern

awk -v r1="2" -v r2="4" ...

2nd line 4th line  
2nd line after second pattern 4th line after second pattern

Using a function

pasteRows(){
  local pattern1="$1"
  local pattern2="$2"
  local frow="$3"
  local srow="$4"

  awk -v pat1="$pattern1" -v pat2="$pattern2" -v r1="$frow" -v r2="$srow" '
   {    
        if($0 == pat1 || $0 == pat2) {i=0; p=$0; next}
        i++; a[p][i] = $0
   } 
   END{for(e in a) print a[e][r1],a[e][r2]}
  ' pattern.txt
}

$ pasteRows pattern "pattern 2" 3 5
3rd line 5th line
3rd line after second pattern 5th line after second pattern
$ pasteRows pattern "pattern 2" 1 4
1st line 4th line  
1st line after second pattern 4th line after second pattern

AWK to find desire lines and merge them

6 Answers6