Replace substring with SED

marri · 07-07-2005, 04:36 AM

hi
I'm trying to figure out how I could replace part of the XML schema I have here below using SED:

Code:

<release>
<supported>
<machine> 213.220.81.240
<type> 103
<ipboot> 225.1.10.1:1103
<ipupgrade> 225.1.10.1:2103
</type>
</machine>
</supported>
<server> 213.220.81.240
<dhcp> 213.220.81.240
<model> 103
<version_active> 0.8.0
<DI> 12
<dir> /test/103/0.8.0
<date> Thu-Jan-20-16:36:51-GMT-2005
</model>
</server>
</release>

I'm trying to use SED to replace the <model> - </model> part with another data, without to much effort. I know I can do this by copying parts to a temp file and play around with in such a way but I was wondering if I could do this with a single line using SED.

Example of <model> - </model> replacement could be:

Code:

<model> 110
<version_active> 0.8.2
<DI> 14
<dir> /test/110/0.8.2
<date> Thu-Jul-07-16:36:51-GMT-2005
</model>

regards
Marri

Chrax · 07-08-2005, 01:44 PM

I don't understand exactly what you're trying to do. Where is the source of the changes you want to make. In this example, how should we know that you want to change

Code:

<model> 103

to

Code:

<model> 110

?

Or do you mean something like:

Code:

sed -i "s/<model> 103/<model> 110/"

but extended to perform the entire switch? If that's the case, sed is hardly an excellent way to do this, as it's just as easy to change the file by hand (if not more so, as you don't have to worry about sed syntax).

eddiebaby1023 · 07-09-2005, 05:18 PM

Use awk. When you find the first pattern (<model>) set a flag that you can test to skip/replace the following lines. When you find the second pattern (</model>) unset the flag. Something like:

Code:

awk '{/a == 1/	{ next }
/<\/model>/	{a=0}
/<model/		{a=1}'  filename

The syntax may need work, I haven't used awk for a while.