UNIX for Dummies Questions & Answers

HI!
I have a news article who I want to tag dates in. One line in the article can for example be "... in the 1800-century it..." and in this text I want to tag the dates with sed like this "... in the <date>1800-century</date> it..." but I don't really get it. A guy helped me with this code:

Code:

sed -e "s/\(1800-century\|1890\|year 1890\|1800 century\)/<date>\1<\/date>/g" article.txt

And he told me that it works fine in Linux but I use Unix and in Unix nothing happened when I tried the codeline, why doesn't it work in Unix?
Thankful for your help.

Unix sed (not GNU sed) usually doesn't support extended regular expression and things like the "or" sequence in your sed sample. Sed DOES take multiple expressions portably, so you could do:

Code:

sed -e 's/\([0-9][0-9][0-9]*-century\)/<date>\1<\/date>/g \
   -e 's/\(year [0-9][0-9][0-9]*\)/<date>\1<\/date>/g \
   -e 's/\([0-9][0-9][0-9]* century\)/<date>\1<\/date>/g'

Depending on the replacement string and further patterns, you can introduce an overlap where a pattern is matched in a subsequent expression. No problem in this case above, but for your 4digit year pattern, if you keep that one, You'll have to do some extra work... (details left out for now).