CODEWITHSUNDEEP
regexsubstringleading-zero
BenTreeser
BenTreeser

Reputation: 5

Regex remove leading zeros if string is number

How can I remove leading zeros if the input is a number and otherwise not?

Specific scenario: I am only interested in the last 10 characters of the input. If this substring is consisting only of numbers, I want to remove the leading zeros. Otherwise, so if there is a word character or a special character, I need this whole substring.

Example 1

input: aaaa000002d111

expected: 000002d111 (because of the 'd' in the substring)

Example 2

input: aaaa0000011111

expected: 11111

I managed to remove the leading zeros with 0*(.{0,10})$ but how do I proceed in case any non-digit is included?

If possible, it would be perfect, if this expected substring is in group 1 of the match.

Upvotes: 0

Views: 4540

Answers (4)

garyh
garyh

Reputation: 2852

.{4}(?:0*([1-9]{1,10}|\w{10}))$

skip the first 4 chars then search for either 10 digits (excluding any leading zeros) or the last 10 characters if not all digits. Both return in Group 1

see demo here

UPDATE

as per comments, changed the regex to allow 0 within the digits (after the first non-zero digit)

.{4}(?:0*([1-9]\d{1,9}|\w{10}))$

see updated demo here

Upvotes: 1

user557597
user557597

Reputation:

This is one way

Find (?=\d{10}$)0*(\d+)$
Replace $1

Says remove leading zeros of last 10 digits.
Note that if the last 10 digits are all 0's, it will leave 1 behind.

Also, the assertion fixes the match position to the last 10 characters
so no need to worry about ranges after that.

Formatted 

 (?= \d{10} $ )        # Assert last 10 are digits
 0*                    # Leading 0's
 ( \d+ )               # (1), The rest
 $                     # EOS

Upvotes: 0

logi-kal
logi-kal

Reputation: 7880

I think this will work:

(?=.{10}$)0*?([1-9][0-9]{0,9}|(?![0-9]*$).*)$

or

(?=.{10}$)0*?([1-9][0-9]{0,9}|.*[^0-9].*)$

First check that we are parsing the last 10 characters ((?=.{10}$)).

Then lazily search for consecutive 0s, followed by a non-0 digit and other (at most 9) digits, or keep everything if there is at least a non-digit character.

In both the cases the output is stored in group 1, as required. See demo.

Upvotes: 0

sfgroups
sfgroups

Reputation: 19099

In Perl you can try this.

cat a |perl -npe 's/^[a-z]+[0]+([1-9][0-9]+)$|^[a-z]+([0]+[\d]+\w\d+)$/\1\2/'

Upvotes: 0

Related Questions