How to identify exact text and then add a blank line above it using sed?

Shell Programming and Scripting
1

I need to identify the exact text of San Antonio Generator Running in the output my script which lands to a text file. Once SED finds the specific text, I need it to insert one line above the matched text.

Here is what I have so far that isn't working all that well for me. Any help would be much appreciated.

sed -i '/"\<San Antonio Generators Running\>/i \\' outagereport.txt
2

Well, you are not searching the pattern that you cite in your text; there's an extra " and an extra s in the pattern. And, try \n for the replacement string.

1 Like
3

What output do you get?

I'm not a sed expert at all, but it would seem that you need to substitute your string with new-line and then your string. I think you are adding literally nothing with your code above.

Could you try something like this? (no promises at all):-

sed 's/"\<San Antonio Generators Running\>"/^J"\<San Antonio Generators Running\>"/' outagereport.txt

The ^J is a single character generated by pressing Ctrl+v then Ctrl+j

I hope that this helps, apologies if it does not.

Robin

1 Like
4

Traditional sed needs a multi-line string:

sed '/\<San Antonio Generators Running\>/ i\
hello world' outagereport.txt

Or

sed '
/\<San Antonio Generators Running\>/ i\
hello world
' outagereport.txt

GNU sed can continue on the same line:

sed '/\<San Antonio Generators Running\>/ ihello world' outagereport.txt

---------- Post updated at 11:11 AM ---------- Previous update was at 11:05 AM ----------

Just seeing that most GNU sed need the traditional approach when inserting a blank line:

sed '/\<San Antonio Generators Running\>/ i\
' outagereport.txt
1 Like
5

You see, I said I didn't quite know what I was doing. :rolleyes:

Thanks MadeInGermany for the corrections. :b:

Robin

6

Okay so I tried most of these and the following worked best for me:

sed '/\<San Antonio Generators Running\>/ i\\' outagereport.txt

This is what the output looked like prior to the SED command:

HOSTNAME USERNAME> cat outagereport.txt
San Antonio Generators Running
San Antonio Generator Alarms: Zero 
San Antonio Battery Discharge
San Antonio Battery Discharge Zero 
San Antonio Comm Power Fail
San Antonio Comm Power Fail Zero 
San Antonio Isolated Sites
San Antonio Isolated Sites Zero

This is the result with the SED command:

HOSTNAME USERNAME> sed '/\<San Antonio Generators Running\>/ i\\' outagereport.txt

San Antonio Generators Running
San Antonio Generator Alarms: Zero 
San Antonio Battery Discharge
San Antonio Battery Discharge Zero 
San Antonio Comm Power Fail
San Antonio Comm Power Fail Zero 
San Antonio Isolated Sites
San Antonio Isolated Sites Zero

Since some three of these strings have the same sequence, such as isolated sites and isolated sites Zero, how can I add a line above only the first instance?

7

sed does not have variables; in order to avoid the match in the further lines one must read them in a loop.
In traditional multi-line syntax (not possible in csh/tcsh):

sed '
/\<San Antonio Generators Running\>/ {
i\

:L
N
$!bL
}
' outagereport.txt

awk variables make things simpler:

awk '(!met && /\<San Antonio Generators Running\>/) {met=1; print ""} {print}' outagereport.txt
1 Like
8

Wouldn't it be simpler to just match entire lines? Using a BSD based sed on Mac OS X with your sample input file:

$ sed '/^San Antonio Battery Discharge$/i\
\

/^San Antonio Comm Power Fail$/i\
\

/^San Antonio Generators Running$/i\
\

/^San Antonio Isolated Sites$/i\
\

' outagereport.txt

San Antonio Generators Running
San Antonio Generator Alarms: Zero 

San Antonio Battery Discharge
San Antonio Battery Discharge Zero 

San Antonio Comm Power Fail
San Antonio Comm Power Fail Zero 

San Antonio Isolated Sites
San Antonio Isolated Sites Zero
$