How to Split OData multi-level expand query string in C#?

Question

I have a URL: Expand=User($select=Firstname,Lastname),Organisation,Contract($Expand=MyOrganisation($select=Name,Status),Organisation),List

I need to split this string in the below format:

User($select=Firstname,Lastname)
Organisation
Contract($Expand=MyOrganisation($select=Name,Status),Organisation)
List

How to achieve this functionality in C#?

Do you have any reasons to have a single parameter instead of 4? Like user=User($select=Firstname,Lastname)&organisation=Organisation, etc. — Boris Sokolov, Jun 11 '21 at 07:38
OData is the thing here "$select=Firstname,Lastname" and it doesn't require to put everything in one string — Boris Sokolov, Jun 11 '21 at 07:53
Odata supports multiple level filters and expand functions Ex: https://stackoverflow.com/questions/51525409/odata-multiple-expands-and-filter — Suresh Sankar, Jun 11 '21 at 08:06

Boris Sokolov · Answer 1 · 2021-06-11T09:11:52.267

More likely you have to use an ODataLib with in-built URI Parser

Uri requestUri = new Uri("Products?$select=ID&$expand=ProductDetail" +
                         "&$filter=Categories/any(d:d/ID%20gt%201)&$orderby=ID%20desc" +
                         "&$top=1&$count=true&$search=tom",
                         UriKind.Relative);
ODataUriParser parser = new ODataUriParser(model, serviceRoot, requestUri);
SelectExpandClause expand = parser.ParseSelectAndExpand(); // parse $select, $expand
FilterClause filter = parser.ParseFilter();                // parse $filter
OrderByClause orderby = parser.ParseOrderBy();             // parse $orderby
SearchClause search = parser.ParseSearch();                // parse $search
long? top = parser.ParseTop();                             // parse $top
long? skip = parser.ParseSkip();                           // parse $skip
bool? count = parser.ParseCount();                         // parse $count

Adding the RegExp option (the fixed version of what Amal has provided below)

    string url = "User($select=Firstname,Lastname),Organisation,Contract($Expand=MyOrganisation($select=Name,Status),Organisation),List";

    Regex rgx = new Regex(@"(.+?)(?:(\(.*?\)),|,)");

    foreach (var match in rgx.Matches($"{url},"))
    {
        Console.WriteLine(match.ToString()[..^1]);
    }

we don't have a model we are using dynamic or Jobject. I was tried it's not working without a model. — Suresh Sankar, Jun 11 '21 at 08:23
A manual parsing would be problematic as you don't have a specific delimiter (comma is used for multiple purposes). So if you know for sure that there is ")," statement, then "," and ")," afterwards - you can go with IndexOf() and substrigns — Boris Sokolov, Jun 11 '21 at 08:45
That's a nice way of using indices and ranges. Note that they are only supported since C# 8.0 — Amal K, Jun 11 '21 at 09:21

Amal K · Accepted Answer · 2021-06-11T16:09:58.570

2

You not only need to split the string but also keep track of the parentheses while splitting. This is not possible with just plain old regex. See this post. However, the splitting can be achieved with some advanced RegEx; .NET fortunately supports balancing groups using which you can keep track of the parentheses. This answer was quite helpful in coming up with a solution. For readability, I have split the regex into multiple lines and used RegexOptions.IgnorePatternWhitespace:

string url = "User($select=Firstname,Lastname),Organisation,Contract($Expand=MyOrganisation($select=Name,Status),Organisation),List";
Regex rgx = new Regex(
               @"(.+?)
                (
                    (
                        \(
                            (?:
                                [^()]
                                |
                                (?'open'\()
                                |
                                (?'close-open'\))
                            )+
                            (?(open)(?!))
                        \)
                    )
                    ,
                    |
                    ,
                    |
                    \b$
                )", 
               RegexOptions.IgnorePatternWhitespace);

foreach(var match in rgx.Matches(url))
{
    Console.WriteLine($"{match.Groups[1]} {match.Groups[3]}"); 
}

The field will be available as match.Groups[1] and the parameters, if any will be available as match.Groups[3](this will be an empty string if there are no parameters). You can access match.Groups[0] to get the entire group.

Regex Breakdown

Regex	Description
`(.+?)`	Non-greedily match one or more characters
`$` and `$`	Match an actual `(` and `)`
`[^()]`	Match any character that is not a `(` or `)`
`(?'open'\()`	Create a named group with the name "open" and match a `(` character
`(?'close-open\))`	Create a group "close" and assign the interval between "open" and "close" to "close" and delete group "open"
`(?(open)(?!))`	Assert if the "open" group is not deleted
`(?:[^()]\|(?'open'$)\|(?'close-open'$))+`	Create a non-capturing group and match one or more characters that match one of the expressions between `\|`

edited Jun 11 '21 at 16:09

answered Jun 11 '21 at 08:48

Amal K

4,359
2
22
44

List will be not part of result and match.Groups[2] will crash – Boris Sokolov Jun 11 '21 at 09:02
@BorisSokolov Thanks, I've found a fix to add `\b$` to match the last word. Also `match.Groups[2]` won't crash, it would just be an empty group. If you check the value of `match.Groups.Count`, it is always 3 since there are two groups in the pattern. – Amal K Jun 11 '21 at 09:18
@SureshSankar that's because how I've formatted it: Change the `Console.WriteLine` to `Console.WriteLine($"{match.Groups[0]}{match.Groups[1]}"` to get the format in the question. – Amal K Jun 11 '21 at 10:10
@SureshSankar Check the updated answer. And specify what exactly isn't working correctly – Amal K Jun 11 '21 at 10:15
@AmalK Console.WriteLine($"{match.Groups[0]}{match.Groups[1]}") i was changed and verify still its not correct – Suresh Sankar Jun 11 '21 at 10:17
1

Expected: Field: Contract Parameters: $Expand=MyOrganisation($select=Name,Status),Organisation) – Suresh Sankar Jun 11 '21 at 10:26
1

@AmalK In contract string split not expected please give me the idea. – Suresh Sankar Jun 11 '21 at 10:27
@SureshSankar Please check the updated answer. – Amal K Jun 11 '21 at 16:10
1

Prefect RegEx explanation :) – Boris Sokolov Jun 11 '21 at 20:22
1

@AmalK and Boris Sokolov That's working fine. Thanks a lot for your supports. – Suresh Sankar Jun 14 '21 at 06:42

Regex	Description
`(.+?)`	Non-greedily match one or more characters
`\(` and `\)`	Match an actual `(` and `)`
`[^()]`	Match any character that is not a `(` or `)`
`(?'open'\()`	Create a named group with the name "open" and match a `(` character
`(?'close-open\))`	Create a group "close" and assign the interval between "open" and "close" to "close" and delete group "open"
`(?(open)(?!))`	Assert if the "open" group is not deleted
`(?:[^()]\|(?'open'\()\|(?'close-open'\)))+`	Create a non-capturing group and match one or more characters that match one of the expressions between `\|`

How to Split OData multi-level expand query string in C#?

2 Answers2

Regex Breakdown