sed on OS X - can't seem to use + in regexps

Question

Now according to all the literature

echo 1234abcd|sed "s|[0-9]\+|#|g"

should return #abcd. And

echo abcd|sed "s|[0-9]\+|#|g"

should return abcd.

But on OS X 10.4.11 the first expression returns 1234abcd. Using * instead of + works for the first example but fails on the second, returning #abcd, because the [0-9] pattern is matched zero times.

Does the + operator not work in regular expressions in OS X? Is there an alternative?

Thanks

Answer 1

Obsolete basic regular expressions do not support + and ? quantifiers. They are regular characters.

Alternatives for [0-9]+ are e.g. [0-9]{1,} or [0-9][0-9]*.

Or you can use sed -E to use modern, extended regular expressions.

Answer 2

On OSX, sed by default uses basic REs. You should use sed -E if you want to use modern REs, including the "+" one-or-more operator.

See here for the indication that sed uses basic REs by default, here for the modern RE syntax, and here for the basic RE (ed) information.

Answer 3

you can use awk

# echo 1234abcd| awk '{gsub(/[0-9]+/,"#")}1'
#abcd

# echo abcd| awk '{gsub(/[0-9]+/,"#")}1'
abcd

Answer 4

Many of the OS X unix utilities are of versions that lack the comforts of their GNU equivalents. As Pax says, you can use -E:

drigz@mbp drigz 0$ echo 1234abcd | /usr/bin/sed "s/[0-9]\+/#/g" 
1234abcd
drigz@mbp drigz 0$ echo 1234abcd | /usr/bin/sed -E "s/[0-9]+/#/g" 
#abcd

Note that small changes to the syntax of your regex are required (\+ to + in this case).

However, I prefer to use fink to get GNU utilities:

drigz@mbp drigz 0$ echo 1234abcd | /sw/bin/sed "s/[0-9]\+/#/g"
#abcd
drigz@mbp drigz 0$ /sw/bin/sed --version
GNU sed version 4.1.5
Copyright (C) 2003 Free Software Foundation, Inc.
This is free software; see the source for copying conditions.  There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE,
to the extent permitted by law.

Answer 5

If + doesn't work, you can always use {1,}