Question

I’m trying to split the following string "Name=='mynme' && CurrentTime<'2012-04-20 19:45:45'" into this:

Name
==
'myname'
&&
CurrentTime
<
'2012-04-20 19:45:45'

I have the following regex:

([+\\-*/%()]{1}|[=<>!]{1,2}|[&|]{2})

The problem is when using the above regex I get the following result:

Name
== 
'myname'
&&
CurrentTime 
<
'2012
-
04
-
20
19:45:45'

I practically need the regex to be quote aware.

Thanks

Update 1 regarding lordcheeto’s answer:

Your response is close. But the following is still not split correctly:

 string input2 = "((1==2) && 2-1==1) || 3+1==4 && Name=='Stefan+123'";

What I need to do is to split a string into operators and operands. Something like this:

 LeftOperand Operator RightOperand

Now, if any operator is between '' it should be ignored and the whole string between '' should be treated as an operand.

The string above should generate the following output:

(

(
1
==
2
)

&&
2
-
1
==
1
)

||
3
+
1
==
4
&&
Name
==
'Stefan+123'

Answer 1

Ok, assuming you want it to simply split on logical and relational operators, you can use this pattern:

string lordcheeto = @"\s*(==|&&|<=|>=|<|>)\s*";    

This will also trim all whitespace from the returned strings.

Code:

using System;
using System.Text.RegularExpressions;

namespace RegEx
{
    class Program
    {
        static void Main(string[] args)
        {
            string original = "([+\\-*/%()]{1}|[=<>!]{1,2}|[&|]{2})";
            string lordcheeto = @"\s*(==|&&|<=|>=|<|>)\s*";

            string input = "Name=='mynme' && CurrentTime<45 - 4";
            string input1 = "Name=='mynme' && CurrentTime<'2012-04-20 19:45:45'";
            string ridiculous = "Name == BLAH && !@#>=$%^&*()< ASDF &&    this          >          that";

            executePattern("original", input, original);
            executePattern("lordcheeto's", input, lordcheeto);
            executePattern("original", input1, original);
            executePattern("lordcheeto's", input1, lordcheeto);
            executePattern("original", ridiculous, original);
            executePattern("lordcheeto's", ridiculous, lordcheeto);
        }

        static void executePattern(string version, string input, string pattern)
        {
            // Avoiding repitition for this example.
            Console.WriteLine("Using {0} pattern:", version);

            // Needs to be trimmed.
            var result = Regex.Split(input.Trim(), pattern);

            // Pipes included to highlight whitespace trimming.
            foreach (var m in result)
                Console.WriteLine("|{0}|", m);

            // Extra space.
            Console.WriteLine();
            Console.WriteLine();
        }
    }
}

Test:

http://goo.gl/XAm6J

Output:

Using original pattern:
|Name|
|==|
|'mynme' |
|&&|
| CurrentTime|
|<|
|45 |
|-|
| 4|


Using lordcheeto's pattern:
|Name|
|==|
|'mynme'|
|&&|
|CurrentTime|
|<|
|45 - 4|


Using original pattern:
|Name|
|==|
|'mynme' |
|&&|
| CurrentTime|
|<|
|'2012|
|-|
|04|
|-|
|20 19:45:45'|


Using lordcheeto's pattern:
|Name|
|==|
|'mynme'|
|&&|
|CurrentTime|
|<|
|'2012-04-20 19:45:45'|


Using original pattern:
|Name |
|==|
| BLAH |
|&&|
| |
|!|
|@#|
|>=|
|$|
|%|
|^&|
|*|
||
|(|
||
|)|
||
|<|
| ASDF |
|&&|
|    this          |
|>|
|          that|


Using lordcheeto's pattern:
|Name|
|==|
|BLAH|
|&&|
|!@#|
|>=|
|$%^&*()|
|<|
|ASDF|
|&&|
|this|
|>|
|that|

---

Edit

Ok, with the additional constraints, you should be able to use this:

string lordcheeto = @"\s*('.*?'|&&|==|<=|>=|<|>|\(|\)|\+|-|\|\|)\s*";

This will still trim all whitespace from the returned strings. It will, however, return empty strings if matches are right next to each other (e.g. Name=='Stefan+123'). I was unable to work around that this time, but it’s not so important.

If you import System.Linq and System.Collections.Generic and make the results a List<string>, you can remove all empty strings from the List in one extra line like this (which is slower than using straight-up for loops):

var results = Regex.Split(input.Trim(), pattern).ToList();
results.RemoveAll(x => x == "");

Code:

using System;
using System.Text.RegularExpressions;

namespace RegEx
{
    class Program
    {
        static void Main(string[] args)
        {
            string lordcheeto = @"\s*('.*?'|&&|==|<=|>=|<|>|\(|\)|\+|-|\|\|)\s*";

            string input = "Name=='mynme' && CurrentTime<45 - 4";
            string input1 = "Name=='mynme' && CurrentTime<'2012-04-20 19:45:45'";
            string input2 = "((1==2) && 2-1==1) || 3+1==4 && Name=='Stefan+123'";

            executePattern("lordcheeto's", input, lordcheeto);
            executePattern("lordcheeto's", input1, lordcheeto);
            executePattern("lordcheeto's", input2, lordcheeto);

            Console.ReadLine();
        }

        static void executePattern(string version, string input, string pattern)
        {
            // Avoiding repitition for this example.
            Console.WriteLine("Using {0} pattern:", version);

            // Needs to be trimmed.
            var result = Regex.Split(input.Trim(), pattern);

            // Pipe included to highlight empty strings.
            foreach (var m in result)
                Console.WriteLine("|{0}", m);

            // Extra space.
            Console.WriteLine();
            Console.WriteLine();
        }
    }
}

Test:

http://goo.gl/lkaoM

Output:

Using lordcheeto's pattern:
|Name
|==
|
|'mynme'
|
|&&
|CurrentTime
|<
|45
|-
|4


Using lordcheeto's pattern:
|Name
|==
|
|'mynme'
|
|&&
|CurrentTime
|<
|
|'2012-04-20 19:45:45'
|


Using lordcheeto's pattern:
|
|(
|
|(
|1
|==
|2
|)
|
|&&
|2
|-
|1
|==
|1
|)
|
|||
|3
|+
|1
|==
|4
|&&
|Name
|==
|
|'Stefan+123'
|

Additional Comments:

If you want to split on any other operators (e.g., <<, +=,=, -=, >>) as well (there’s a lot), or need anything else, just ask.