Bashscripts.org

Hi!

I have some xml-files without reference to a dtd and css. I want to add such a reference, and thought of a script like:

#!/bin/sh
for i in *.xml
do sed {s/<?xml version="1.0" encoding="ISO-8859-1"?>/`<?xml version="1.0" encoding="ISO-8859-1"?>\n<!DOCTYPE inventory SYSTEM "file.dtd">\n<?xml-stylesheet type="text/css" href="file.css"?>`}
done

(Replacing the first line of each file with itself and two other lines.)
However, I get an error that the map or file doesn't exist.
What would be the right syntaxis?

Well sed is not built to operate without input data...

This is one method


Another is


Both these methods will just output the replaced lines. The second method can be altered to change the files on-the-fly by adding -i instead of -e or to create a new file (still using -e) and just redirecting output to a new filename (> tmp2.txt)

Best regards
Fredrik Eriksson

Fredrik, thanks!

I found it difficult to figure out how to interpret "test data", "test \(data\)" or "\1", but I also found out that I was too much orientated at a bash script that I (with some help) had written almost two years ago, while this could be done with a single command.

In the end, the command looked like:

Sorry about that. It's like second nature for me so I miss to explain it more in-depth sometimes

"test data" is just some random text string to run with sed.
\(data\) is a regexp buffer command. It will save "data" from the string matched ("test data" in this case).
and \1 is just the position of the buffer. If you add 2 \(<something>\) then the first is \1 and the second is \2

Best regards
Fredrik Eriksson