Regex usage in grep

Question

Im trying to search for a pattern in a file as follows:

SRC_ERROR_CODE=105
SRC_ERROR_CODE=106
...

To achieve this following is the grep statement used:

grep -io "[a-z]*_error_code=[0-9]*" events.log

However i was wondering if instead of using the "*" which fetches 0 to n occurrences of the preceding matched character, the "+" should fetch the same results as well as below:

grep -io "[a-z]+_error_code=[0-9]+" events.log

But, this doesn't seem to work. Could you please guide as to why it doesn't.

Thanks

Answer 1

The + in regex matches one or more character in Extended Regular Expression (ERE - this is egrep)

In Basic regular expression (BRE) you have to escape the + sign.

Since you are using grep, you have to escape the +. Therefore, use \+ instead of +.

If you are using egrep, you can use the unescaped +

Answer 2

In POSIX Basic Regular Expressions (BRE), the default regex dialect used by grep, + just matches itself.

In POSIX Extended Regular Expressions (ERE) and Perl Compatible RegEx (PCRE), + matches 1 or more of the preceding atom.

You can ask grep to use ERE with the -E option:

$ echo "foo baaar" | grep -o -E 'a+'
aaa