Regex: Match everything except backreference

Question

I have the following example lines:

a_a
b_c

How (using grep / egrep ) would I match the lines where the first letter is not equal to the last letter? I have tried the following but this does not seem to work correctly.

egrep ([ab])_[^\1]

Working with egrep -v or the inverse regex, (match everything except where first letter equals last letter), is not possible for my use case.

Answer 1

Use group matches.

There you create a group with the 1st character, and then put group ( \\1 ) at the end.

^(.).*\\1$

egrep '^(.).*\1$'  

Opposite:

^(.).*((?!\\1).)+$

pcregrep '^(.).*((?!\1).)+$'

Answer 2

May be you are looking for this:

\\b([az])\\w+(?!(\\1))([az])\\b

it works for:

a_a and b_c

(tested with: The Regulator 2.0.3)

You can then adjust this regex to meet further your needs

Answer 3

Use a negative look behind anchored to end of input. A general regex for "first char is not last char" is:

^(.).*(?<!\1)$

To match only your type of input:

^(.)_.(?<!\1)$