Reading numbers from string in C#

Question

What I want?

I want to display weather information on my page. I want to display the result in the browser specific culture.

What am I doing?

I use MSN RSS for this purpose. MSN returns the report in XML format. I parse the XML and display results.

What problem am I facing?

When displaying the report, I have to parse an XML node, <data> which will be different values in different culture.

For e.g.,

en-US: "Lo: 46°F. Hi: 67°F. Chance of precipitation: 20%"

de-DE: "Niedrig: 46°F. Höchst: 67°F. Niederschlag %: 20%"

I want to read only low, high and chance of precipitation values. i.e., I want to read 46, 67 and 20%.

Can somebody please give me a solution for this?

May be RegX or someother method is also fine with me :-)

Thanks in advance!

You get temperature in °F even for a German locale? Makes me wonder where's the point in having all this culture-specific, actually ... — Joey
– Joey, Commented Sep 21, 2009 at 10:54
Sorry, the URL I used had a query string requesting the report to be in F :). The URL I used is this: weather.msn.com/… — Vijay
– Vijay, Commented Sep 21, 2009 at 11:02
Ok, since you stick the culture in the query string anyway, why do you even do that? If you are only interested in the numbers, then why even bother with other locales you won't even ever need? — Joey
– Joey, Commented Sep 21, 2009 at 11:07
@Johannes Rössel: Actually, I should not be sticking the culture in query string. That is the mistake I have done when forming the question. :) — Vijay
– Vijay, Commented Sep 21, 2009 at 11:18

Tor Haugen · Accepted Answer · 2009-09-21 11:30:16Z

3

You should consider always fetching the RSS using the same culture. That way, you'll have an easier task parsing the content. If you'll only be using the numbers, it shouldn't stop you from emitting culture-specific content to the end user.

So if you go for the en-US version, you could do it like this:

Regex re = new Regex(@"Lo: (\d+)°F. Hi: (\d+)°F. Chance of precipitation: (\d+)%");
var match = re.Match(forecast);
if (match.Success)
{
    var groups = match.Groups;
    lo = int.Parse(groups[1].Captures[0].Value);
    hi = int.Parse(groups[2].Captures[0].Value);
    prec = int.Parse(groups[3].Captures[0].Value);
}

answered Sep 21, 2009 at 11:30

Tor Haugen

19.7k9 gold badges48 silver badges64 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

Vijay Over a year ago

I combined your and Johannes Rössel's answers to get my problem solved. Thanks!

Joey · Accepted Answer · 2009-09-21 10:58:59Z

2

If you only want the numbers, you can use a regular expression, for example the following:

(\d+).*?(\d+).*?(\d+%)

A quick test in PowerShell shows that it does work at least for your input data:

PS Home:\> function test ($re) {
>>   $a -match $re; $Matches
>>   $b -match $re; $Matches
>> }
>>
PS Home:\> $a = "Lo: 46°F. Hi: 67°F. Chance of precipitation: 20%"
PS Home:\> $b = "Niedrig: 46°F. Höchst: 67°F. Niederschlag %: 20%"
PS Home:\> test "(\d+).*?(\d+).*?(\d+%)"
True

Name                           Value
----                           -----
3                              20%
2                              67
1                              46
0                              46°F. Hi: 67°F. Chance of precipitation: 20%
True
3                              20%
2                              67
1                              46
0                              46°F. Höchst: 67°F. Niederschlag %: 20%

However, it won't work anymore if any locale might use numbers in the description strings.

You can add other constraints, like requiring a colon before every match:

: (\d+).*?: (\d+).*?: (\d+%)

This should deal with spurious numbers elsewhere in the string. But the best way overall would actually be to get your data from a source which gives you the data for machine reading, not for human consumption

answered Sep 21, 2009 at 10:58

Joey

357k88 gold badges704 silver badges699 bronze badges

1 Comment

Vijay Over a year ago

The RegeX worked. I combined your and Tor Haugen's answers to get my problem solved. Thanks!

ICR · Accepted Answer · 2009-09-21 11:30:19Z

1

The following should extract the two numbers and chance of precipitation, as well as the units that are used (for culturally dependent units).

(?<lo>\d+°.).*?(?<hi>\d+°.).*?(?<precipitation>\d+)

If you don't want units extracted, then you can use

(?<lo>\d+)°.*?(?<hi>\d+)°.*?(?<precipitation>\d+)

answered Sep 21, 2009 at 11:30

ICR

14.2k4 gold badges52 silver badges79 bronze badges

Comments

RvdK · Accepted Answer · 2009-09-21 10:59:16Z

0

use regex (but i don't know the regex formula ;) )

You can also do a forloop over the sentence, and check each char if it's a integer. Each time you encounter once, place it in a string. when finding something else than an integer, parse the string to an int and voila. Do this 3 times

answered Sep 21, 2009 at 10:59

RvdK

19.8k5 gold badges68 silver badges110 bronze badges

4 Comments

Vijay Over a year ago

Certainly! I can do this. But, is that the best solution?

RvdK Over a year ago

I would do a regex like johannes. definately cleaner but harder to read.

RvdK Over a year ago

By the way, there's no best solution. Every solution has its pro and cons. regex pro: clean, smaller and probably faster. Regex con: harder to read, hard to master regex

Vijay Over a year ago

Ok. Let me check it out. Thanks :)

Mike Nowak · Accepted Answer · 2009-09-21 11:18:22Z

Its quite weird you are not getting XML with values in different nodes which would make more sense to me (they you could pick which values use for different locales).

But, if you want to extract data from given strings try this or something simmilar if you are not a fan of RegEx:

string dataUS = "Lo: 46°F. Hi: 67°F. Chance of precipitation: 20%";
string dataDE = "Niedrig: 46°F. Höchst: 67°F. Niederschlag %: 20%";
string[] stringValues = dataU.Split(new string[] {": "}, 4, StringSplitOptions.None);
List<int> values = new List<int>();
for (int i = 1; i < 4; i++)
{
    StringBuilder sb = new StringBuilder();
    foreach (char c in stringValues[i].Trim())
    {
        if (Char.IsDigit(c))
        {
            sb.Append(c);
        }
        else
        {
            values.Add(Convert.ToInt32(sb.ToString()));
            break;
        }
    }
}

(im spliting on ": " instead of digits)

Beatles1692 · Accepted Answer · 2009-09-21 11:21:21Z

0

I suggest using Regex to get the values that you want according to UI culture language one by one : I mean you can have a Regex to get the Lo temp. "(Lo|Niedrig):(\d+)" , a regex to get Hi temp "(Hi|Höchst):(\d+)" and a regex to get chance of perception and so on. In all of the above examples you can get the number from second element of the match.

answered Sep 21, 2009 at 11:21

Beatles1692

5,35040 silver badges70 bronze badges

1 Comment

Joey Over a year ago

You can also use non-grouping parentheses for the literal parts: (?:Lo|Niedrig) to avoid having groups you don't do anything with.

Collectives™ on Stack Overflow

Reading numbers from string in C#

6 Answers 6

1 Comment

1 Comment

Comments

4 Comments

Comments

1 Comment

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

6 Answers 6

1 Comment

1 Comment

Comments

4 Comments

Comments

1 Comment

Your Answer

Sign up or log in

Post as a guest

Linked

Related