regex for numbers with possible digits after dot or comma in text

I need a regex that can filter out numbers that might have digits after the dot from a text.

(?:[A-Z]\s)(\d+)((.|,)(\d+))?

Text : Expected results : Actual Results
H 24    : 24            : 24
24.5    : 24.5          : 24
24,5    : 24.5          : 24
test5   : 5             : 5
test 5.5: 5.5           : 5
50.752  : 50.752        : 50

For some reason PG only returns the first group of my match. Can someone help me please.

select substring('test 5.5', cast('(?:[A-Z]*\s*)(\d+)((.|,)(\d+))?' as character varying)) as convertedvalue

You could use a combination of substring() and replace(), like:

select x, replace(substring(x, '(\d+(.|,)*\d*)'), ',', '.') convertedvalue

substring() captures the relevant part of the string, then replace() replaces the , with . if needed.

Demo on DB Fiddle

with t(x) as ( values ('H 24'), ('24.5'), ('24,5'), ('test5'), ('test 5.5'), ('50.752') )
select x, replace(substring(x, '(\d+(.|,)*\d*)'), ',', '.') convertedvalue
from t

x        | convertedvalue
:------- | :-------------
H 24     | 24            
24.5     | 24.5          
24,5     | 24.5          
test5    | 5             
test 5.5 | 5.5           
50.752   | 50.752