Get floating number match on regex using python

use Lookahead after the match to locate the float number before string GB/Day(case insensitive): (?= GB/Day)

[\d.]+(?= GB/Day|GB/Day| gb/day|gb/day)

Regex101 Demo

Here is a fix to the immediate problem with your pattern:

input = "Roaming Calls, 1.5 GB/Day 100 SMS/Day"
m0 = re.match(r'.*?([0-9.]+)?(gb|GB| gb| GB)', input)
if m0:
    print "match: ", m0.group(1)

Just make the dot appearing right before the capture group for the number lazy.

Demo

For both float and other numbers you can try this :

import re
k = "Roaming Calls, 1.5 GB/Day 100 SMS/Day"
print(re.findall(r"[-+]?\d*\.\d+|\d+",k))

if you want to find only float values go for this :

import re
k = "Roaming Calls, 1.5 GB/Day 100 SMS/Day"
print(re.findall(r"[-+]?\d*\.\d+",k))

it will return a list of float numbers in that string like this :

['1.5']

The issue that .* matches everything and leaves only 1 symbol for [0-9.]+. You can replace it with .? so it won't be that greedy:

.?([0-9.]+)(gb|GB| gb| GB)

regex101