Filter with * which will restrain all values that starts with Null?

Question

I have the following code which extends my filter functionality to allow me to search for more than one Zip-code. Ex. (1254,125,145,1455)

Current Code:

if (Post_Code_PAR != null)
{
    String[] PostCode_Array = Post_Code_PAR.Split(',');
    query = from p in query where PostCode_Array.Contains(p.Post_Code) select p;
}

I want to extend this code in a way that if I give it something like (0*) it will find zip codes which starts with 0. If I give it 123* it will give me all zip codes which starts with 123. Thanks a lot.

Do you also want to be able to do tings like `*123` for things that end with `123` or `1*1` for things that ends and starts with `1`? — Hjalmar Z, Apr 15 '15 at 11:09

jeromerg · Answer 1 · 2015-04-15T13:34:36.180

3

use regular expression:

example:

IEnumerable<string> allInputs = new[] {"70000", "89000", "89001", "90001"};
string commaSeparatedPostCodePatterns = @"89000 , 90\d\d\d";

if (string.IsNullOrWhiteSpace(commaSeparatedPostCodePatterns)) 
    return;

string[] postCodePatterns = commaSeparatedPostCodePatterns
                                .Split(',')
                                .Select(p => p.Trim()) // remove white spaces
                                .ToArray();

var allMatches = 
    allInputs.Where(input => postCodePatterns.Any(pattern => Regex.IsMatch(input, pattern)));

foreach (var match in allMatches)
    Console.WriteLine(match);

With such a problem, the initial requirements are very simple but quickly become more and more complex (due to internationalization, exceptions to the rule, some smart tester testing an unexpected limit-case like 90§$% ...). So regular expressions provide the best trade-off simplicity vs. extendability

The regex equivalent to * is .*. But in your case, you would rather need more restrictive patterns and use the \d placeholder matching a single digit. Optional digit would be \d?. Zero or more digits would be \d*. One or more digits would be \d+. Please look at the documentation for more details.

edited Apr 15 '15 at 13:34

answered Apr 15 '15 at 11:25

jeromerg

2,997
2
26
36

I just used `Console.WriteLine()` to check that it works in a sample program: it writes into the dos console the result. You can ignore the last two lines. Remark: do prefer the linq syntax `myList.Where(...)` instead of `from myItem in myList ...`: In most case, it is more readable. – jeromerg Apr 15 '15 at 12:54
see the addition to the response – jeromerg Apr 15 '15 at 13:04
If you want to keep the single `* ` in your post code patterns, you can sustitute it on the fly at a the beginning with `myString.Replace("*", ".*")`. That way you keep the power of regex for future extensions and continue avoiding `if` statement. – jeromerg Apr 15 '15 at 13:16
You can copy&paste the sample code into the `public static void Main()` entry point of a C# console application, It compiles and runs. And from this point then you can adapt it to your need and test the result. – jeromerg Apr 15 '15 at 13:19
Remark: I'm not a regex expert at all. I've just learned to use them a few years ago for a similar need. The best way to learn, that I found, was by using the regexbuddy software. But you can also ask on stackoverflow, maybe in a separate question. – jeromerg Apr 15 '15 at 13:25
@Zayed: the `Split()` doesn't remove the spaces after/before the comma. Try removing them. You can remove them systematically by using `.Trim()`. I will add it to the sample. – jeromerg Apr 15 '15 at 13:32
One last question in which part of your code Can I user my Parameter ? "Post_code_PAR"? coz you didnit use it at all . I should change the second line to [string commaSeparatedPostCodePatterns = Post_Code_PAR;] – Dragon Apr 15 '15 at 13:40
can you check the first answer plz, its the simplest ? – Dragon Apr 15 '15 at 13:48
@Zayed: sql `like` is similar to regex in c#. The sample you provided contains two important components: at first, you look for multiple post code patterns (`PostCode_Array`) and second, the search must be performed with wildcards. In SQL it wouldn't be straightforward neither: as you should combine both `ANY` and `LIKE` operators and use a sub-select. – jeromerg Apr 15 '15 at 13:48
@Zayed: `PostCode_Array` corresponds to `postCodePatterns` and contains the list of patterns your are looking for. `query` corresponds to `allInputs` and contains the list of objects to filter (in your example containing the property `Post_Code`, and in my sample being the postCode itself) – jeromerg Apr 15 '15 at 14:02

Hjalmar Z · Answer 2 · 2015-04-15T14:33:36.890

3

You can replace the * with \d* which is the Regex way of saying "any amount of digits", basically. Then run through and filter with Regex.IsMatch().

if (Post_Code_PAR != null)
{
    var PostCode_Array = from p in Post_Code_PAR.Split(',') select String.Format("^{0}$", p.Replace("*", @"\d*"));
    query = (from p in query where PostCode_Array.Any(pattern => Regex.IsMatch(p.Post_Code, pattern)) select p).ToArray();
}

I tried to keep it as close to your code as possible but you could clean it up a bit and use lambda instead. But this should work just as well :)

edited Apr 15 '15 at 14:33

answered Apr 15 '15 at 11:47

Hjalmar Z

1,591
1
18
36

Add `using System.Text.RegularExpressions;` at the top of the document. – Hjalmar Z Apr 15 '15 at 11:57
Where does it say that? – Hjalmar Z Apr 15 '15 at 12:01
FYI:the field post code at my data set is a string and the parameter too its a string – Dragon Apr 15 '15 at 12:03
I think I forgot to get the property, try again – Hjalmar Z Apr 15 '15 at 12:06
Hi now I have a red line under your "select" with this error "Error 9 The type 'System.Collections.Generic.IEnumerable "can not be implicitly converted to' string [] '. There is already an explicit conversion exists. (Maybe you missing a cast.) " – Dragon Apr 15 '15 at 12:08
Sorry still not working, can I send you a screen shot by mail ? – Dragon Apr 15 '15 at 12:43
@Zayed Saw your mail, try my updated answer. If that doesn't work try putting it within parenthesis and put .ToArray(9 at the end like I did on the next line. – Hjalmar Z Apr 15 '15 at 14:35
@ Hjalmar: Thanks, now the code is error free but the result set is returning all lines in the table. as if there is not filter – Dragon Apr 15 '15 at 15:21
@ Hjalmar: I sent you a pic with the results :) – Dragon Apr 15 '15 at 15:27
@Zayed Can you step through the code with a break-point and see if everything is running through? – Hjalmar Z Apr 15 '15 at 15:48

Bondaryuk Vladimir · Answer 3 · 2015-04-15T14:36:49.857

2

string Post_Code_PAR = "";
String[] PostCode_Array = Post_Code_PAR.Split(',');
var zipCodeList = query.Where(x => CustomFilter(x.Post_Code, PostCode_Array)).ToList();

private static bool CustomFilter(string code, string[] foundZipCodes)
{
    return foundZipCodes.Any(x =>
   {
       x = x.Trim();
       var text = x.Replace("*", "");
       if (x.EndsWith("*"))
           return code.StartsWith(text);
       if (x.StartsWith("*"))
           return code.EndsWith(text);
       if (x.Contains("*"))
           return false;
       return code.StartsWith(text);
    });

}

now user can find: "0*,*52" etc.

edited Apr 15 '15 at 14:36

answered Apr 15 '15 at 11:06

Bondaryuk Vladimir

509
3
10

Hi Bondaryuk: what about this line "query = from p in query where PostCode_Array.Contains(p.Post_Code) select p;" Should I delete it ? – Dragon Apr 15 '15 at 11:32
@Zayed yes, you can. if you want only startwith use var zipCodeList = PostCode_Array.Where(x => x.StartsWith(startWith)); – Bondaryuk Vladimir Apr 15 '15 at 11:38
Hi, I donot get what you do this var filter = "*0";?. I have a Parameter called "Post_Code_PAR" which the user enter vales in it. I wanted to filter those vales in case he wrote some thing like 1455* -- > this will get me all vales which start with 1455 – Dragon Apr 15 '15 at 11:51
@Zayed sorry, i understood you, now I change answer, now user can found: "*231, 254*" it return wher StartsWith and EndWith – Bondaryuk Vladimir Apr 15 '15 at 13:19
you mean if i enter 123* i will get all number which start with 123? coz this is what I want – Dragon Apr 15 '15 at 14:11
@Zayed oh... it's my bug, i changed filter – Bondaryuk Vladimir Apr 15 '15 at 14:34
@Zayed _you mean if i enter 123* i will get all number which start with 123?_ - yes you are right – Bondaryuk Vladimir Apr 15 '15 at 15:51

Filter with * which will restrain all values that starts with Null?

3 Answers3