how to delete text from line starting pattern1 up to line before pattern2?

Shell Programming and Scripting
1

My data is xml'ish (here is an excerpt) :-

  <bag name="mybag1" version="1.0"/>
  <contents id="coins"/>
  <bag name="mybag2" version="1.1"/>
  <contents id="clothes"/>
  <contents id="shoes"/>
  <bag name="mybag3" version="1.6"/>

I want to delete line containing mybag2 and its subsequent contents (number of contents lines can vary). Thus I wish to delete from pattern mybag2 up to (but not including) the next "bag name" tag and result in :-

  <bag name="mybag1" version="1.0"/>
  <contents id="coins"/>
  <bag name="mybag3" version="1.6"/>

I have tried this a few different ways with sed and awk and have yet to find a solution. Any help would be appreciated.

2

This is my solution:

file : awk_tets

BEGIN { flag = 0}
/^<bag name="mybag2"/ { flag = 1}
/^<bag name="mybag3"/ { flag = 0}
{ if (flag == 0) { print; } }

in command line:

$ awk -f awk_test your_data_file > result

Let's try it :slight_smile:

3

this is what i had in mind too, problem is that it may not handel nested ceses too well.....

4

I tried that solution :-

me@myserver $ nawk ' BEGIN { flag = 0} /^<bag name="mybag2"/ { flag = 1} /^<bag name="mybag3"/ { flag = 0} { if (flag == 0) { print; }}' test2

<bag name="mybag1" version="1.0"/>
<contents id="coins"/>
<bag name="mybag3" version="1.6"/>

It works, but the issue is that the second bag name could be any value, its not specifically "mybag3", so the second pattern must be the more generic "/^bag name=/

So i tried that also :-

me@myserver $ nawk ' BEGIN { flag = 0} /^<bag name="mybag2"/ { flag = 1} /^<bag name=/ { flag = 0} { if (flag == 0) { print; }}' test2

<bag name="mybag1" version="1.0"/>
<contents id="coins"/>
<bag name="mybag2" version="1.1"/>
<contents id="clothes"/>
<contents id="shoes"/>
<bag name="mybag3" version="1.6"/>

it failed because the flag was being reset straight after it was set as the more generic pattern also matched the mybag2 line.

Then I switched the pattern order and BINGO !

me@myserver  $ nawk ' BEGIN { flag = 0} /^<bag name=/ { flag = 0} /^<bag name="mybag2"/ { flag = 1} { if (flag == 0) { print; }}' test2<bag 

name="mybag1" version="1.0"/>
<contents id="coins"/>
<bag name="mybag3" version="1.6"/>

Thanks very much for leading me in the right direction kholostoi

5

Or:

awk '/^<bag name=/{f=0}$0~v{f=1}!f' v="mybag2" file

Use nawk or /usr/xpg4/bin/awk on Solaris.

6

much more elegant, thanks radoulov

I will use that