Why is Regex Group Extraction Giving Wrong Answer?

Question

I am trying to get the Number for an Integer number (No Decimal exist), or the number part before the Decimal Point for a Decimal Number.

The example below should give 12345 as the answer. But it is giving it as 1234 (eating the 5 out)

            string isDecimalTypeNumbers = @"^(?<wholeNumberPart>[0-9]*)\.?[0-9]+?$";
            Regex pattern = new Regex(isDecimalTypeNumbers);
            Match match = pattern.Match("12345");
            if (match.Success)
            {
                string wholeNumberPartWithoutPlusMinusSign = match.Groups["wholeNumberPart"].Value;
                MessageBox.Show(wholeNumberPartWithoutPlusMinusSign);
            }

It is giving 1234 please try it yourself. I need the integer number if it is an integer number, or the integer part for a decimal number. — BeeGees, Feb 01 '20 at 06:26
My bad! I misunderstood the question at first. Use `^(?[0-9]*)(?:\.[0-9]+)?$` instead. This makes the whole "decimal part" optional. — 41686d6564 stands w. Palestine, Feb 01 '20 at 06:29
@AhmedAbdelhameed you are correct. I need to group the decimal part as that is also optional. — BeeGees, Feb 01 '20 at 06:35
What is the ?: in (?:\.[0-9]+) when we already have a ? at the end for the entire bracket (?:\.[0-9]+)? — BeeGees, Feb 01 '20 at 07:03
It's called a [non-capturing group](https://stackoverflow.com/q/3512471/8967612). — 41686d6564 stands w. Palestine, Feb 01 '20 at 07:09
Thanx Ahmed. One more important thing - Why does Regex give Match Success if Value String is Empty? — BeeGees, Feb 01 '20 at 09:13
Just tried an empty string and, for me, `Match match = pattern.Match("");` yields `match.Success == False`. — AdrianHHH, Feb 01 '20 at 09:54
If you don't want an empty string to match, try `^(?[0-9]+)(?:\.[0-9]+)?$` https://regex101.com/r/yPR3lK/1 — The fourth bird, Feb 01 '20 at 10:19

score 1 · Answer 1 · answered Feb 01 '20 at 06:31

1

^(?<wholeNumberPart>[0-9]*)

captures the first part of the number

\.?

captures the decimal point if it exists

[0-9]+?$

captures the last number regardless of whether there is a decimal point

So as you can see the last bit captures the 5 leaving only 1234 as the wholeNumberPart.

I think this explains why the regex group extraction is giving you the wrong answer of 1234

answered Feb 01 '20 at 06:31

openshac

4,966
5
46
77

thanx shac. But still why should (?[0-9]*) not capture the last digit 5, because after all 12345 is a whole number without any decimals. – BeeGees Feb 01 '20 at 06:37
2

If `[0-9]*` were to match `12345` then there'd be no digits left for `[0-9]+?` to match. As [@Xiaoy312's answer](https://stackoverflow.com/a/60014622/150605) says, `+?` is lazy but not optional, so `[0-9]+?` must match at least one digit. – Lance U. Matthews Feb 01 '20 at 06:45
Got it @bacon. Your explanation makes it clearer. Thanx. – BeeGees Feb 01 '20 at 06:56
Why does Regex give Match Success if Value String is Empty? – BeeGees Feb 01 '20 at 09:14

score 1 · Accepted Answer · answered Feb 01 '20 at 06:32

1

The decimal separator (\.?) is optional, so it may not match.
The fractional part ([0-9]+?) is not optional, but lazy, so it will still try to match.

So, what could happen is 1234 is consumed by the whole number part, and \.? doesnt match, but 5 is consumed by the fractional part.

You should make the decimal separator and the fractional part into one optional group:

@"^(?<wholeNumberPart>[0-9]*)(\.[0-9]+)?$"

answered Feb 01 '20 at 06:32

Xiaoy312

14,292
1
32
44

Thanx Xiaoy. I tried this (the same as @AhmedAbdelhameed recommended) and it worked. Thanx. – BeeGees Feb 01 '20 at 06:43
Why does Regex give Match Success if Value String is Empty? – BeeGees Feb 01 '20 at 09:14
Just change the `[0-9]*` to `[0-9]+`. The `*` matches 0-to-many elements, so nothing is also a valid match. Replacing it with `+` will make it matches 1-to-many elements. – Xiaoy312 Feb 01 '20 at 18:05

Why is Regex Group Extraction Giving Wrong Answer?

2 Answers2