How to extract a line and a line before that based on certain strings

Question

I am trying to extract lines containing a certain string also 4 lines before it (just the fourth line not the ones in between). Right now I can get the line itself;

grep -n "my string" myfile.out > output.txt

I can also get the lines before it, but not that specific one, all the four preceding lines;

grep "my string" -B 4 myfile.out > output.txt

Let's take this example:

 ! CREEK                                                              
 FLUX 6739, 6670, 6640, 6672, 6582, 6583, 6548, 6576, 6570 &      
 656565                   
 ***
 ***  THE SUM OF SPECIFIED:   -1.192451    
 ***  THE AVERAGE:  -6.2760599E-02
 ***  THE MAXIMUM:    0.000000    
 ***  THE MINIMUM:  -0.4294980    
 ***  THE SUM OF POSITIVE:    0.000000    
 ***  THE SUM OF NEGATIVE:   -1.192451

I want this output:

 ! CREEK
 ***  THE SUM OF SPECIFIED:   -1.192451

I would appreciate windows-batch-file solution or PowerShell as well.

Here's [an answer](https://stackoverflow.com/questions/16588795/powershell-how-to-obtain-previous-lines-from-text-file) that illustrates a PowerShell way to accomplish this — BenH, Jun 08 '17 at 14:43

score 2 · Answer 1 · answered Jun 08 '17 at 16:18

Poweshell solution :

# verbose version 
Select-String "THE SUM OF SPECIFIED" -Path "C:\temp\test.txt" -Context 4,0 | foreach { $_.Context.PreContext[0]; $_.Line}

# or short version (with alias)
sls "THE SUM OF SPECIFIED" "C:\temp\test.txt" -Co 4,0 | % { $_.Context.PreContext[0]; $_.Line}

score 1 · Accepted Answer · answered Jun 08 '17 at 14:46

1

Here is tac + awk solution:

tac file |
awk '/THE SUM OF SPECIFIED/{p=NR; print} p && NR==p+4{print; p=0}' |
tac

! CREEK
***  THE SUM OF SPECIFIED:   -1.192451

Alternatively here is grep + awk solution:

grep "THE SUM OF SPECIFIED" -B 4 file | awk 'NR==1 || NR ==5'

answered Jun 08 '17 at 14:46

anubhava

761,203
64
569
643

1

I should accept your answer and I will. But a follow up question that I don't wanna post another question for it. What if I want to find the first line before `"The sum of ..."` that start with **`!`**. Cause I realized sometimes it is third line before that. – M-- Jun 08 '17 at 14:53
1

For that you can do: `tac file | awk '/THE SUM OF SPECIFIED/{p=NR; print} p && /^!/{print; p=0}' | tac` – anubhava Jun 08 '17 at 14:56

score 0 · Answer 3 · answered Jun 08 '17 at 14:41

0

If your file is named myfile.out this should do what you want to:

grep "THE SUM OF SPECIFIED" -B 4  myfile.out | sed -n "1p;5p"

answered Jun 08 '17 at 14:41

xiawi

1,772
4
19
21

It doesn't do anything. – M-- Jun 08 '17 at 14:45
the `sed` part should display the first and firth lines only... Works on classical `bash`. @anubhava has just give an equivalent solution with `awk` – xiawi Jun 08 '17 at 14:49
I am on windows using cygwin. Will give it a try when I got home Thanks.\ – M-- Jun 08 '17 at 15:00

How to extract a line and a line before that based on certain strings

3 Answers3