Shell Programming and Scripting

Hi, I'm very new to shell scripting and have searched google and this forum for quite some time now.

I have the following in my xml file:

Code:

<recipients>
        <member>value1</member>
</recipients>

I need to find a string <recipients> that follows with a new-line and bunch of spaces and <member>blah blah, i.e.

Code:

<recipients>
        <member>value1</member>

And I need to replace it in the same format, i.e.

Code:

<recipients>
        <member>NEWVALUE</member>

There are exactly 8 spaces in front of the <member> tag but I'd like to avoid doing matching on # of spaces.

Here's what I've tried so far:

Code:

sed 's#<recipients.*#<member>NEWVALUE</member>#' file.xml

the final string looks like this:

Code:

....
       <member>NEWVALUE</member>
</recipients>

--the opening tag <recipients> gets deleted, which makes sense based on what i put in the sed command, I guess.

Code:

sed 's#<recipients.*#<recipients><member>NEWVALUE</member>#' file.xml

--the final string looks like this:

Code:

<recipients><member>NEWVALUE</member> -- new line inserted as a result of sed 
        <member>%%adminemail%%</member> -- old line remained in the file
</recipients>

-- again, makes sense based on the sed command but I just don't know how I can preserve the format of the string and look for <recipients> \n spaces <member> and replace it with the same string , just with different value.

I have originally tried using '/' instead of # as a seprator char, but switched to # becuase the value1 I'm looking to replace may contain a bunch of special chars, including '/' slash.

At this point I'm not even sure this can be done with sed? Should I try using "awk" or even perl, maybe?

Any help is greatly appreciated!

Hello,
I don't know if the next is a solution...

Code:

tratar_linea=0  # false
echo "" > $new_file

while read line
do
   new_line="$line"
   if [ $tratar_linea -eq 1 ]
   then
       echo $line | grep '<tag2>' > /dev/null
       if [ $? -eq 0 ]
       then
          new_line="<tag2>new value you want</tag2>"
       fi
   fi

   echo $line | grep '<tag1>' > /dev/null
   if [ $? -eq 0 ]
   then
      tratar_linea=1
   else
      tratar_linea=0
   fi

   echo $new_line >> $new_file
done < xml_file_you_have

The following approach is fine if you are going to hardcode "NEWVALUE" into the replacement text.

Code:

$ cat data
<recipients>
        <member>value1</member>
<recipients>
<member>value2</member>
$ sed '/<recipients>/{N;s/<member>[^<]*/<member>NEWVALUE/;}' data
<recipients>
        <member>NEWVALUE</member>
<recipients>
<member>NEWVALUE</member>

If instead it needs to be parameterized (if so, you should always mention that when requesting assistance), then you would need to make sure that NEWVALUE cannot include characters which are special in replacement text (such as the s command's delimiter, &, \, etc). In such a case, it would probably be safest to use AWK and its string handling functions (index, substr, length, and their ilk).

Regards,
Alister

Beautiful, as all solutions in this forum....

I didn't know 'N' in sed, I suppose is the Next line? or new line character?

Hi, alberto:

sed's N command appends a newline followed by the next line in the input to the line currently in the pattern space.

Regards,
Alister

yes, indeed, beatiful and is working great!!!

Thank you alister. I don't need to parameterize the replacement string so this solution is exactly what I was looking for. And thanks for clarification on N.

albertogarcia - thanks for the reply as well!

-Marina