parsing logic proving very complex

Question

I have a string that i must parse, setting a classes vars correctly from it. The string is badly structured but i can not change it. I have tried parsing it but dont have a good way to do it without issue. The string itself is a set of attributes and params that are used to trigger a exe from cmd line.

I have laid it out in a way to make easier to read but know that is it 1 continuous string.

Here are the rules as to how to read this line. For every 'non -dll' command you can expect a single key and value pair. for the dlls lines you can have a single or multiple key-value pairs after the initial dll=,' ie the 'dll' element contains 0 or more keyValue or stand alone value split by spaces. eg dll=oneMoreDll, andItsParam=value anotherParam=value lastParam=value value

Input string

time=value1 size=value2 dll=aDllName dll=anotherDllName, someParam=ParamValue dll=yetAnotherDll, someOhterParam=anotherParamValue aStandAloneValue dll=oneMoreDll, andItsParam=value anotherParam=value lastParam=value

I want to be able to parse this string into the following format, i was thinking each line in a string array.

I have tried splitting by spaces and then 'dll' but my regex aint up to scratch or its impossible (im sure its not). Help!

Desired output elements, to be stored in a String array

time=value1 
size=value2 
dll=aDllName 
dll=anotherDllName, someParam=ParamValue  
dll=yetAnotherDll, someOhterParam=anotherParamValue aStandAloneValue
dll=oneMoreDll, andItsParam=value anotherParam=value lastParam=value

Answer 1

The following should work, at least for the sample case.

1. Split the string by ' '
2. Split each sub-string by '='. If there's no '=', we simply take the left side.

We're now left with a structure that looks something like this:

{ left = attribute1, right = value1 }, { left = attribute2, right = value2 }, { left = aStandAloneValue }, etc.

Now, we need to group each item by the previous 'dll'. I'm using an extension method taken from this answer to help with that.

Essentially, it will group until the condition is not met. In our case, we want to stop grouping when we hit a 'dll' entry. Or, if we haven't yet hit a 'dll' entry, then we always create a new group.

The rest is simply formatting the output (which may not be needed in your case).

var inStr = "time=value1 size=value2 dll=aDllName dll=anotherDllName, someParam=ParamValue dll=yetAnotherDll, someOhterParam=anotherParamValue aStandAloneValue dll=oneMoreDll, andItsParam=value anotherParam=value lastParam=value";

bool isBeforeAnyDll = true;

var result = inStr.Split(new[] { ' ' }, StringSplitOptions.RemoveEmptyEntries)
    .Select(r => {
        var split = r.Split('=');
        if (split.Length == 1)
            return new { left = split[0], right = (string)null };

        var left = split[0];
        var right = split[1];
        return new { left, right };
    })
    .GroupAdjacentBy((l, r) =>  {
        return r.left == "dll" 
             ? isBeforeAnyDll = false
             : !isBeforeAnyDll;
    })
    .Select(g => string.Join(" ", 
        g.Select(gg => { 
            if (gg.right == null)
                return gg.left;
            return string.Format("{0}={1}", gg.left, gg.right);
        })));


//https://stackoverflow.com/a/4682163/563532
public static class LinqExtensions
{
    public static IEnumerable<IEnumerable<T>> GroupAdjacentBy<T>(
        this IEnumerable<T> source, Func<T, T, bool> predicate)
    {
        using (var e = source.GetEnumerator())
        {
            if (e.MoveNext())
            {
                var list = new List<T> { e.Current };
                var pred = e.Current;
                while (e.MoveNext())
                {
                    if (predicate(pred, e.Current))
                    {
                        list.Add(e.Current);
                    }
                    else
                    {
                        yield return list;
                        list = new List<T> { e.Current };
                    }
                    pred = e.Current;
                }
                yield return list;
            }
        }
    }
}

Output:

time=value1 
size=value2 
dll=aDllName 
dll=anotherDllName, someParam=ParamValue 
dll=yetAnotherDll, someOhterParam=anotherParamValue aStandAloneValue 
dll=oneMoreDll, andItsParam=value anotherParam=value lastParam=value 

The data is all properly grouped together after the .GroupAdjacentBy(), the following code is simply formatting the output.

Answer 2

Why not split by Enviroment.NewLine then by , then split by the first = sign, take the left part as the variable name, then the right part as the variable value?

Answer 3

You can use the following regex approach using Regex.Matches:

using System;
using System.Linq;
using System.Text.RegularExpressions;
public class Test
{
    public static void Main()
    {
        var log = "time=value1 size=value2 dll=aDllName dll=anotherDllName, someParam=ParamValue dll=yetAnotherDll, someOhterParam=anotherParamValue aStandAloneValue dll=oneMoreDll, andItsParam=value anotherParam=value lastParam=value";
        var res = Regex.Matches(log, @"\bdll=(?:(?!\bdll=).)*|\w+=\w+")
                 .Cast<Match>()
                 .Select(p => p.Value)
                 .ToList();
        Console.WriteLine(string.Join("\n",res));
    }
}

See IDEONE demo and a regex demo

The regex matches 2 alternatives:

• \bdll= - a whole word dll= followed with...
• (?:(?!\bdll=).)* - zero or more characters that are not dll
• | - or....
• \w+=\w+ - one or more word characters followed with = followed with one or more word characters.