UNIX for Dummies Questions & Answers

I'm trying to use grep on XML files. The same grep expressions work on plaint text files but not on XML files (which of course are plain text files). Actually, these expressions work on XML files saved in DreamWeaver, but not when the same files are saved in XML Spy.

I want grep to treat these files as plain text which is what I assumed it would do.

Any pointers?

Thanks,

Paul

I don't know about Spy's output. First thing: can you vi those files and read them?
If you are not getting text behavior in vi, then spy may be doing something interesting.

Or try:
diff spy.xml dm.xml

Where these two (spy.xml and dm.xml) should be identical. They probably are not.

My wild guess would be that XML Spy is saving them in something crazy like 16-bit Unicode.

Thanks for your input.

Corona688, you said "My wild guess would be that XML Spy is saving them in something crazy like 16-bit Unicode" and I think that's it. (Don't know how to test it though?

After further experimentation, I've discovered that this only happens when I save it encoded as UTF-16 (i.e. the declaration is <?xml version="1.0" encoding="UTF-16"?>).

When Encoding = UTF-16
Saved in DreamWeaver - file is accessible to grep
Saved in XmlSpy - not accessible to grep

When Encoding = UTF-8
Saved in DreamWeaver - accessible to grep
Saved in XmlSpy - accessible to grep

So if XmlSpy is saving them in 16-bit unicode, can I use grep with them, or do I need to turn to something else?

Cheers

If you're saving it as UTF-16, you're saving it as 16-bit Unicode, and my guess was right on the money. You can't grep UTF-16. It uses 16-bit characters instead of the normal 8 bits. Grep is comparing half a character of UTF-16 to one full character of some other set. Naturally it won't work.

Can you save it as UTF-8 instead? It can represent all UNICODE characters, without breaking grep.