Parse time string using regex - regex

Question

My time string may be in one of the following formates (x and y - integer numbers, h and m - symbols):
xh ym
xh
ym
y
Examples:
1h 20m
45m
2h
120
What regular expression should I write to get x and y numbers from such string?

Answer 1

(\d+)([mh]?)(?:\s+(\d+)m)?
You can then inspect groups 1-3. For your examples those would be
('1', 'h', '20')
('45', 'm', '')
('2', 'h', '')
('120', '', '')
As always, you might want to use some anchors ^, $, \b...

Answer 2

I'm going to assume you're using .NET due to your username. :)
I think in this case, it's easier to use TimeSpan.ParseExact for this task.
You can specify a list of permitted formats (see here for the format for these) and ParseExact will read in the TimeSpan according to them.
Here is an example:
var formats = new[]{"h'h'", "h'h 'm'm'", "m'm'", "%m"};
// I have assumed that a single number means minutes
foreach (var item in new[]{"23","1h 45m","1h","45m"})
{
TimeSpan timespan;
if (TimeSpan.TryParseExact(item, formats, CultureInfo.InvariantCulture, out timespan))
{
// valid
Console.WriteLine(timespan);
}
}
Output:
00:23:00
01:45:00
01:00:00
00:45:00
The only problem with this is that it is rather inflexible. Additional whitespace in the middle will fail to validate. A more robust solution using Regex is:
var items = new[]{"23","1h 45m", "45m", "1h", "1h 45", "1h 45", "1h45m"};
foreach (var item in items)
{
var match = Regex.Match(item, #"^(?=\d)((?<hours>\d+)h)?\s*((?<minutes>\d+)m?)?$", RegexOptions.ExplicitCapture);
if (match.Success)
{
int hours;
int.TryParse(match.Groups["hours"].Value, out hours); // hours == 0 on failure
int minutes;
int.TryParse(match.Groups["minutes"].Value, out minutes);
Console.WriteLine(new TimeSpan(0, hours, minutes, 0));
}
}
Breakdown of the regex:
^ - start of string
(?=\d) - must start with a digit (do this because both parts are marked optional, but we want to make sure at least one is present)
((?<hours>\d+)h)? - hours (optional, capture into named group)
\s* - whitespace (optional)
((?<minutes>\d+)m?)? - minutes (optional, capture into named group, the 'm' is optional too)
$ - end of string

Answer 3

I would say that mhyfritz' solution is simple, efficient and good if your input is only what you shown.
If you ever need to handle corner cases, you can use a more discriminative expression:
^(\d+)(?:(h)(?:\s+(\d+)(m))?|(m?))$
But it can be overkill...
(get rid of ^ and $ if you need to detect such pattern in a larger body of text, of course).

Answer 4

Try this one: ^(?:(\d+)h\s*)?(?:(\d+)m?)?$:
var s = new[] { "1h 20m", "45m", "2h", "120", "1m 20m" };
foreach (var ss in s)
{
var m = Regex.Match(ss, #"^(?:(\d+)h\s*)?(?:(\d+)m?)?$");
int hour = m.Groups[1].Value == "" ? 0 : int.Parse(m.Groups[1].Value);
int min = m.Groups[2].Value == "" ? 0 : int.Parse(m.Groups[2].Value);
if (hour != 0 || min != 0)
Console.WriteLine("Hours: " + hour + ", Mins: " + min);
else
Console.WriteLine("No match!");
}

Answer 5

in bash
echo $string | awk '{for(i=1;i<=NF;i++) print $i}' | sed s/[hm]/""/g