How can I get RegEx.Split to extract a series of numbers containing decimal points from a string?

Question

This sort of example for extracting integers from a string is common:

string input = "10 blah 20 30 nonsense 40 50";

string[] numbers = System.Text.RegularExpressions.Regex.Split(input, @"^[\d]");

But how can I account for numbers with a decimal point?

e.g.

string input = "10 blah 20 30 nonsense 40.5 50"

used with the above regular expression, unsurprisingly, the 40 and the 5 after the decimal point are split into different elements of the numbers array.

In my naivity, I thought the below would work:

string[] numbers = System.Text.RegularExpressions.Regex.Split(input, @"^[\d\.]");

But this causes the decimal point to be split into its own element of the array.

This seems like it should be so easy but I've tried all sorts of regular expressions without any success. I'm tearing my hair out - any help is greatly appreciated!

Answer 1

System.Text.RegularExpressions.Regex.Split(input, @"[^\d.]+");

The split delimiter is 1 or more characters that are neither digits nor periods. Note that your common example is wrong even for the first input. It will just use the first digit as the delimiter, which means it doesn't appear in the array.

EDIT: I tested it with both inputs, and it works:

string input = "10 blah 20 30 nonsense 40 50";
System.Text.RegularExpressions.Regex.Split(input, @"[^\d.]+");
{ "10", "20", "30", "40", "50" }

input = "10 blah 20 30 nonsense 40.5 50";
System.Text.RegularExpressions.Regex.Split(input, @"[^\d.]+");
{ "10", "20", "30", "40.5", "50" }

Answer 2

Bear in mind that with Regex.Split(input, @"[^\d.]+");, you are going to get false positives (for strings like non.sense or 50.6.8) => you will have to filter out the results with the value '.' in them.

You could also use the Matches method instead of Split.

MatchCollection mc = Regex.Matches(input, @"\d+(?:\.\d+)?");<br />
string[] numbers = (from Match m in mc select m.Value).ToArray();