We Keep Coding

Dynamic linq filter children using pivot

I have business objects that look like the following:
class Project
{
public int ID
{
get;set;
}
public string ProjectName
{
get;set;
}
public IList<ProjectTag> ProjectTags
{
get;set;
}
}
class ProjectTag
{
public int ID
{
get;set;
}
public int ProjectID
{
get;set;
}
public string Name
{
get;set;
}
public string Value
{
get;set;
}
}
Example Data:
Project:
ID ProjectName
1 MyProject
ProjectTags:
ID ProjectID Name Value
1 1 Name 1 Value 1
2 1 Name 2 Value 2
3 1 Name 3 Value 3
Basically it's a way for our users to define their own columns on the Project. As a result, it's important to remember that I don't know the names of the ProjectTag entries at design time.
What I'm trying to accomplish is to give our users the ability to select projects based on search criteria using System.Linq.Dynamic. For instance, to select just the project in my example above, our users could enter this:
ProjectName == "MyProject"
The more complicated aspect is applying a filter to the ProjectTags. Our application currently allow users to do this in order to filter Projects by their ProjectTags:
ProjectTags.Any(Name == "Name 1" and Value == "Value 1")
That works, but starts to get a bit messy for end users to use. Ideally I'd like to write something that would let them do the following:
Name 1 == "Value 1"
Or if necessary (due to white space in the name), something like the following...
[Name 1] == "Value 1"
"Name 1" == "Value 1"
For lack of a better explanation, it seems like I want to do the equivalent of a SQL pivot on the ProjectTags, and then still be able to execute a where clause against that. I've looked at some of the questions on StackOverflow about pivots and dynamic pivoting, but I haven't found anything too useful.
I've also been thinking about looping through all the ProjectTag Names and building a dynamic query using a left join on each. I guess something like this:
select
Project.*,
Name1Table.Value [Name 1],
Name2Table.Value [Name 2],
Name3Table.Value [Name 3]
from
Project
left join ProjectTag Name1Table on Name = 'Name 1'
left join ProjectTag Name2Table on Name = 'Name 2'
left join ProjectTag Name3Table on Name = 'Name 3'
And then take that query and apply a where clause to it. But I'm not really sure how to do that in Linq as well as dealing with the white space in the name.
I also came across ExpandoObject. I thought possibly I could convert Project to an ExpandoObject. Then loop through all known ProjectTag names, adding each name to the ExpandoObject and, if that Project had a ProjectTag for that name, use that ProjectTag value as the value, else empty string. For example...
private static object Expand(
Project project,
List<string> projectTagNames)
{
var expando = new ExpandoObject();
var dictionary = (IDictionary<string, object>) expando;
foreach (var property in project.GetType()
.GetProperties())
{
dictionary.Add(property.Name, property.GetValue(project));
}
foreach (var tagName in projectTagNames)
{
var tagValue = project.ProjectTags.SingleOrDefault(p => p.Name.Equals(tagName));
dictionary.Add(tagName, tagValue?.Value ?? "");
}
return expando;
}
The exciting thing about this solution is I have an object that looks exactly like I think it should prior to filtering with a where clause. It even seems to accommodate spaces in the property name.
Then of course I found out that dynamic linq doesn't work nicely with ExpandoObject, and so it can't find the dynamic properties. I guess that's because it essentially has a type of Object which isn't going to define any of the dynamic properties. Maybe it's possible to generate a type at run time that matches? Even if that works, I don't think it can account for spaces in the Name.
Am I trying to accomplish too much with this functionality? Should I just tell the users to use syntax like ProjectTags.Any(Name == "Name1" and Value == "Value1")? Or is there some way to trick dynamic linq into understanding ExpandoObject? Seems like having a way to override the way dynamic linq resolves property names would be very handy.

How about using a translator to convert tag references?
I assume that tag names containing spaces will be surrounded by brackets ([]) and that Project field names are a known list.
public static class TagTranslator {
public static string Replace(this string s, Regex re, string news) => re.Replace(s, news);
public static string Surround(this string src, string beforeandafter) => $"{beforeandafter}{src}{beforeandafter}";
public static string SurroundIfMissing(this string src, string beforeandafter) => (src.StartsWith(beforeandafter) && src.EndsWith(beforeandafter)) ? src : src.Surround(beforeandafter);
public static string Translate(string q) {
var projectFields = new[] { "ID", "ProjectName", "ProjectTags" }.ToHashSet();
var opREStr = #"(?<op>==|!=|<>|<=|>=|<|>)";
var revOps = new[] {
new { Fwd = "==", Rev = "==" },
new { Fwd = "!=", Rev = "!=" },
new { Fwd = "<>", Rev = "<>" },
new { Fwd = "<=", Rev = ">=" },
new { Fwd = ">=", Rev = "<=" },
new { Fwd = "<", Rev = ">" },
new { Fwd = ">", Rev = "<" }
}.ToDictionary(p => p.Fwd, p => p.Rev);
var openRE = new Regex(#"^\[", RegexOptions.Compiled);
var closeRE = new Regex(#"\]$", RegexOptions.Compiled);
var termREStr = #"""[^""]+""|(?:\w|\.)+|\[[^]]+\]";
var term1REStr = $"(?<term1>{termREStr})";
var term2REStr = $"(?<term2>{termREStr})";
var wsREStr = #"\s?";
var exprRE = new Regex($"{term1REStr}{wsREStr}{opREStr}{wsREStr}{term2REStr}", RegexOptions.Compiled);
var tq = exprRE.Replace(q, m => {
var term1 = m.Groups["term1"].Captures[0].Value.Replace(openRE, "").Replace(closeRE, "");
var term1q = term1.SurroundIfMissing("\"");
var term2 = m.Groups["term2"].Captures[0].Value.Replace(openRE, "").Replace(closeRE, "");
var term2q = term2.SurroundIfMissing("\"");
var op = m.Groups["op"].Captures[0].Value;
if (!projectFields.Contains(term1) && !term1.StartsWith("\"")) { // term1 is Name, term2 is Value
return $"ProjectTags.Any(Name == {term1q} && Value {op} {term2})";
}
else if (!projectFields.Contains(term2) && !term2.StartsWith("\"")) { // term2 is Name, term1 is Value
return $"ProjectTags.Any(Name == {term2q} && Value {revOps[op]} {term1})";
}
else
return m.Value;
});
return tq;
}
}
Now you just translate your query:
var q = "ProjectName == \"Project1\" && [Name 1] == \"Value 1\" && [Name 3] == \"Value 3\"";
var tq = TagTranslator.Translate(q);

How to sort strings by a different value

I've tried looking for an existing question but wasn't sure how to phrase this and this retrieved no results anywhere :(
Anyway, I have a class of "Order Items" that has different properties. These order items are for clothing, so they will have a size (string).
Because I am OCD about these sorts of things, I would like to have the elements sorted not by the sizes as alphanumeric values, but by the sizes in a custom order.
I would also like to not have this custom order hard-coded if possible.
To break it down, if I have a list of these order items with a size in each one, like so:
2XL
S
5XL
M
With alphanumeric sorting it would be in this order:
2XL
5XL
M
S
But I would like to sort this list into this order (from smallest size to largest):
S
M
2XL
5XL
The only way I can think of to do this is to have a hard-coded array of the sizes and to sort by their index, then when I need to grab the size value I can grab the size order array[i] value. But, as I said, I would prefer this order not to be hard-coded.
The reason I would like the order to be dynamic is the order items are loaded from files on the hard disk at runtime, and also added/edited/deleted by the user at run-time, and they may contain a size that I haven't hard-coded, for example I could hard code all the way from 10XS to 10XL but if someone adds the size "110cm" (aka a Medium), it will turn up somewhere in the order that I don't want it to, assuming the program doesn't crash and burn.
I can't quite wrap my head around how to do this.

Also, you could create a Dictionary<int, string> and add Key as Ordering order below. Leaving some gaps between Keys to accomodate new sizes for the future. Ex: if you want to add L (Large), you could add a new item as {15, "L"} without breaking the current order.
Dictionary<int, string> mySizes = new Dictionary<int, string> {
{ 20, "2XL" }, { 1, "S" },
{ 30, "5XL" }, { 10, "M" }
};
var sizes = mySizes.OrderBy(s => s.Key)
.Select(s => new {Size = s.Value})
.ToList();

You can use OrderByDescending + ThenByDescending directly:
sizes.OrderByDescending(s => s == "S")
.ThenByDescending( s => s == "M")
.ThenByDescending( s => s == "2XL")
.ThenByDescending( s => s == "5XL")
.ThenBy(s => s);
I use ...Descending since a true is similar to 1 whereas a false is 0.

I would implement IComparer<string> into your own TShirtSizeComparer. You might have to do some regular expressions to get at the values you need.
IComparer<T> is a great interface for any sorting mechanism. A lot of built-in stuff in the .NET framework uses it. It makes the sorting reusable.
I would really suggest parsing the size string into a separate object that has the size number and the size size then sorting with that.

You need to implement the IComparer interface on your class. You can google how to do that as there are many examples out there

you'll have to make a simple parser for this. You can search inside the string for elements like XS XL and cm" if you then filter that out you have your unit. Then you can obtain the integer that is the value. If you have that you can indeed use an IComparer object but it doesn't have that much of an advantage.

I would make a class out of Size, it is likely that you will need to add more functionality to this in the future. I added the full name of the size, but you could also add variables like width and length, and converters for inches or cm.
private void LoadSizes()
{
List<Size> sizes = new List<Size>();
sizes.Add(new Size("2X-Large", "2XL", 3));
sizes.Add(new Size("Small", "S", 1));
sizes.Add(new Size("5X-Large", "5XL", 4));
sizes.Add(new Size("Medium", "M", 2));
List<string> sizesShortNameOrder = sizes.OrderBy(s => s.Order).Select(s => s.ShortName).ToList();
//If you want to use the size class:
//List<Size> sizesOrder = sizes.OrderBy(s => s.Order).ToList();
}
public class Size
{
private string _name;
private string _shortName;
private int _order;
public string Name
{
get { return _name; }
}
public string ShortName
{
get { return _shortName; }
}
public int Order
{
get { return _order; }
}
public Size(string name, string shortName, int order)
{
_name = name;
_shortName = shortName;
_order = order;
}
}

I implemented TShirtSizeComparer with base class Comparer<object>. Of course you have to adjust it to the sizes and objects you have available:
public class TShirtSizeComparer : Comparer<object>
{
// Compares TShirtSizes and orders them by size
public override int Compare(object x, object y)
{
var _sizesInOrder = new List<string> { "None", "XS", "S", "M", "L", "XL", "XXL", "XXXL", "110 cl", "120 cl", "130 cl", "140 cl", "150 cl" };
var indexX = -9999;
var indexY = -9999;
if (x is TShirt)
{
indexX = _sizesInOrder.IndexOf(((TShirt)x).Size);
indexY = _sizesInOrder.IndexOf(((TShirt)y).Size);
}
else if (x is TShirtListViewModel)
{
indexX = _sizesInOrder.IndexOf(((TShirtListViewModel)x).Size);
indexY = _sizesInOrder.IndexOf(((TShirtListViewModel)y).Size);
}
else if (x is MySelectItem)
{
indexX = _sizesInOrder.IndexOf(((MySelectItem)x).Value);
indexY = _sizesInOrder.IndexOf(((MySelectItem)y).Value);
}
if (indexX > -1 && indexY > -1)
{
return indexX.CompareTo(indexY);
}
else if (indexX > -1)
{
return -1;
}
else if (indexY > -1)
{
return 1;
}
else
{
return 0;
}
}
}
To use it you just have a List or whatever your object is and do:
tshirtList.Sort(new TShirtSizeComparer());
The order you have "hard-coded" is prioritized and the rest is put to the back.
I'm sure it can be done a bit smarter and more generalized to avoid hard-coding it all. You could e.g. look for sizes ending with an "S" and then check how many X's (e.g. XXS) or the number before X (e.g. 2XS) and sort by that, and then repeat for "L" and perhaps other "main sizes".

Dynamically Generate Codes in C#

I have a list of Enums like the following:
public enum Evaluation : int
{
//Section 1
S1_1_1 = 579,
S1_1_2 = 584,
S1_1_3 = 589,
S1_1_4 = 594,
S1_1_5 = 599,
S1_1_6 = 604,
//Section 2
S1_2_1 = 610,
S1_2_2 = 615,
S1_2_3 = 620,
S1_2_4 = 625,
S1_2_5 = 630,
};
I want to iterate each section and use the values dynamically
int S1Count = 6;
for (int i = 1; i <= S1Count; i++)
{
VoteCount += string.IsNullOrEmpty(this.GetEvaluationValue(FormID, Evaluation.S1_1_ + i)) ? 0 : 1;
}
How can I achieve that? Thanks.
Sorry, my mistake. I tried to get the value from the database by using enum values which are IDs and I have to calculate counts, average for each section.

You can use Enum.Parse to do what you want I think though I don't reccomend it.
To use enum.Parse you'd just need to do something like:
Enum.Parse(typeof(Evaluation), String.Format("S1_1_{0}",i));
This does point at you using some dodgy methodology though. As I said in comments above you would be better off with a data structure allowing you to have sections and their contents easily differentiated. You can do this with either custom classes or maybe just a dictionary of Lists of ints...
Dictionary<int, List<int>> SectionContents;
and use it like:
foreach(int id in SectionContents[sectionNumber])
{
VoteCount += string.IsNullOrEmpty(this.GetEvaluationValue(FormID, id)) ? 0 : 1;
}
(I don't vouch for what's in the foreach, I'm just demonstrating how a dictionary of a list of ints could work).
Creating the Dictionary is easy enough and doesn't require enums. And if this is database stuff could easily be generated through a database query to get the IDs and what sections they are in and then create the data structure.

This will do it
public class Program
{
public static void Main()
{
foreach (FieldInfo fInfo in typeof(Evaluation).GetFields(BindingFlags.Public | BindingFlags.Static))
{
Console.WriteLine("Evaluation." + fInfo.Name);
}
Console.ReadLine();
}
}

Using strings instead of enums?

Is it common place to use a string for comparison as opposed to an enum?

I am aware about your context, but as a first step you can just refactor this way:
Step 1
if (typeOfObject == "UAV")
{
DoSomeWork(_stkObjectRootToIsolateForUavs);
}
else if (typeOfObject == "Entity")
{
DoSomeWork(_stkObjectRootToIsolateForEntities);
}
private void DoSomeWork(IAgStkObject agStkObject)
{
IAgStkObject stkObject = agStkObject.CurrentScenario.Children[stkObjectName];
IAgDataProviderGroup group = (IAgDataProviderGroup)stkUavObject.DataProviders["Heading"];
IAgDataProvider provider = (IAgDataProvider)group.Group["Fixed"];
IAgDrResult result = ((IAgDataPrvTimeVar)provider).ExecSingle(_stkObjectRootToIsolateForUavs.CurrentTime);
stkObjectHeadingAndVelocity[0] = (double)result.DataSets[1].GetValues().GetValue(0);
stkObjectHeadingAndVelocity[1] = (double)result.DataSets[4].GetValues().GetValue(0);
}
Then consider replasing if's with switch:
Step 2
switch (typeOfObject)
{
case "UAV":
DoSomeWork(_stkObjectRootToIsolateForUavs);
break;
case "Entity":
DoSomeWork(_stkObjectRootToIsolateForEntities);
break;
default:
throw new NotImplementedException():
}
This can be even better when using enums.

At the very least, the strings should be declared as constants (or perhaps readonly fields) somewhere, instead of spread out through the code. However, this looks like the schoolbook example for when to use an enum.
public enum ObjectType
{
UAV,
Entity,
// and so on
}

To add to #Restuta's answer, I'd use a
IDictionary<MyEnumifiedString, Action<IAgStkObject>>
to get rid of that if.

I'd agree with #Frederik that this seems a perfect case for using enums, but it could be that the only thing you can get out of the application is a string. In which case your example is perfectly OK.
Oh yes - and make sure you have the string constants defined in one place, preferably a config file so that if they change the other application you don't have to recompile yours.

Regarding your first question I will always use a defined type to store the strings simply to have one location for change if needed.
So for your example i would have the following
public sealed class RootTypes
{
public const string Entity = "entity";
public const string UAV = "uav";
}
Your code then updates to this
typeOfObject = typeOfObject.ToLower();
if (typeOfObject == RootTypes.UAV)
{
stkUavObject = _stkObjectRootToIsolateForUavs.CurrentScenario.Children[stkObjectName];
var group = (IAgDataProviderGroup) stkUavObject.DataProviders["Heading"];
var provider = (IAgDataProvider) group.Group["Fixed"];
IAgDrResult result = ((IAgDataPrvTimeVar) provider).ExecSingle(_stkObjectRootToIsolateForUavs.CurrentTime);
stkObjectHeadingAndVelocity[0] = (double) result.DataSets[1].GetValues().GetValue(0);
stkObjectHeadingAndVelocity[1] = (double) result.DataSets[4].GetValues().GetValue(0);
}
else if (typeOfObject == RootTypes.Entity)
{
IAgStkObject stkEntityObject = _stkObjectRootToIsolateForEntities.CurrentScenario.Children[stkObjectName];
var group = (IAgDataProviderGroup) stkEntityObject.DataProviders["Heading"];
var provider = (IAgDataProvider) group.Group["Fixed"];
IAgDrResult result = ((IAgDataPrvTimeVar) provider).ExecSingle(_stkObjectRootToIsolateForEntities.CurrentTime);
stkObjectHeadingAndVelocity[0] = (double) result.DataSets[1].GetValues().GetValue(0);
stkObjectHeadingAndVelocity[1] = (double) result.DataSets[4].GetValues().GetValue(0);
}
The issue of code redundancy has been anserwed by Restuta

Use enums with bit flags:
[Flags]
public enum MyFlags
{
SomeFlag = 0x1, // 001
OtherFlag = 0x2,// 010
ThirdFlag = 0x4 // 100
}
var firstObject = MyFlags.SomeFlag;
var secondObject = MyFlags.SomeFlag | MyFlags.OtherFlag;
if(((int)secondObject & MyFlags.SomeFlag) != 0)
{
// true
}
if(((int)secondObject & MyFlags.OtherFlag) != 0)
{
// true
}
if(((int)firstObject & MyFlags.SomeFlag) != 0)
{
// true
}
if(((int)firstObject & MyFlags.OtherFlag) != 0)
{
// false
}
This article would be helpful.

Most efficient way to parse a flagged enum to a list

I have a flagged enum and need to retrieve the names of all values set on it.
I am currently taking advantage of the enum's ToString() method which returns the elements comma-separated.
public void SetRoles(Enums.Roles role)
{
IList<Entities.Role> roleList = role.ToString("G").Split(',')
.Select(r => new Entities.Role(r.Trim()))
.ToList();
...
}
I'm sure there must be a better way than this.

Try this:
public void SetRoles(Enums.Roles role)
{
List<string> result = new List<string>();
foreach(Roles r in Enum.GetValues(typeof(Roles)))
{
if ((role & r) != 0) result.Add(r.ToString());
}
}

If you genuinely just want the strings, can't get much simpler than:
string[] flags = role.ToString().Split(',');
This is simpler than using LINQ and is still just a single line of code.
Or if you want a list instead of an array as in the sample in the question you can convert the array into a list:
List<string> flags = new List<string>(role.ToString().Split(','));
In my case I needed a generic solution and came up with this:
value.ToString().Split(',').Select(flag => (T)Enum.Parse(typeof(T), flag)).ToList();

Enum.Parse will handle the concatenated values outputted by ToString just fine. Proof using the Immediate window:
? System.Enum.Parse(typeof(System.AttributeTargets), "Class, Enum")
Class | Enum
(the second line is the output, which is different in the debugger/immediate window from the generic Enum.ToString() output).

List<string> GetRoleNames(Roles roles) =>
Enum.GetValues(typeof(Roles))
.Cast<Roles>()
.Where(role => roles.HasFlag(role))
.Select(role => role.ToString())
.ToList();
void TestRoleSelection()
{
var selectedRoles = (Roles)6;
var roleNames = GetRoleNames(selectedRoles);
Console.WriteLine(string.Join(",", roleNames));
// Output: Admin,User
}
[Flags]
enum Roles
{
SuperAdmin = 1,
Admin = 2,
User = 4,
Anonymous = 8
}

Why do you need a list? Everything is already stored in the flags:
[Flags]
enum Roles
{
Read = 0x1,
Write = 0x2,
Delete = 0x4,
}
Then assign roles:
var roles = Roles.Read | Roles.Write;
And whenever you need to check if a given role has been you don't need to look in a list, but simply look in the roles enumeration:
if ((roles & Roles.Read) == Roles.Read)
{
// The user has read permission
}
if ((roles & Roles.Write) == Roles.Write)
{
// The user has write permission
}

Similar answer to Mick's but puts the operations into extensions and fixes/cleans up the extra space character (from the split).
Also as a bonus if the enum has a _ in it, the code changes it to a space.
public static class EnumExtensions
{
// Take anded flag enum and extract the cleaned string values.
public static List<string> ToComparableStrings(this Enum eNum)
=> eNum.ToString()
.Split(',')
.Select(str => str.ToCleanString())
.ToList();
// Take an individual enum and report the textual value.
public static string ToComparableString(this Enum eNum)
=> eNum.ToString()
.ToCleanString();
// Remove any spaces due to split and if `_` found change it to space.
public static string ToCleanString(this string str)
=> str.Replace(" ", string.Empty)
.Replace('_', ' ');
}
Usage
var single = PivotFilter.Dollars_Only;
var multiple = PivotFilter.Dollars_Only | PivotFilter.Non_Productive;
// These calls return:
single.ToComparableString() // "Dollars Only"
multiple.ToComparableString() // "Non Productive,Dollars Only"
multiple.ToComparableStrings() // List<string>() { "Non Productive", "Dollars Only" }
Enum for Usage
[Flags]
// Define other methods, classes and namespaces here
public enum PivotFilter
{
Agency = 1,
Regular = 2,
Overtime = 4,
Non_Productive = 8,
Dollars_Only = 16,
Ignore = 32
}

Turning a flagged enum to a list might not be as straight forward as it looks. Take the following enum for example:
[Flags]
enum MenuItems
{
None = 0,
Pizza = 1,
Fries = 2,
Pancakes = 4,
Meatballs = 8,
Pasta = 16,
StuffWithP = Pizza | Pancakes | Pasta,
All = Pizza | Fries | Pancakes | Meatballs | Pasta | StuffWithP
};
If we have the value StuffWithP, what do we want in the list? StuffWithP or Pizza, Pancakes, Pasta? I had a use case in witch I needed to "deconstruct" the enum value to the invidual flags and put those in a list. I came up with the following:
public static string[] DeconstructFlags(Enum items)
{
if (items.GetType().GetCustomAttribute<FlagsAttribute>() == null)
{
throw new ArgumentException("Enum has no [Flags] attribute.", nameof(items));
}
// no value, no list
var itemsValue = (int)(object)items;
if (itemsValue == 0) return Array.Empty<string>();
var result = new List<string>();
foreach (var item in Enum.GetValues(items.GetType()))
{
if(item == null) continue;
var value = (int)item;
// skip combined flags
if (!BitOperations.IsPow2(value))
{
continue;
}
if (items.HasFlag((Enum)item))
{
result.Add(item.ToString() ?? "");
}
}
return result.ToArray();
}
I don't know if it is the most efficient, but it skips those combined flags nicely. Wrote some more on my blog: Deconstructing a [Flags] enum.

F# version
module Enum =
let inline flagsToList< ^e when ^e: equality and ^e: (static member (&&&): ^e * ^e -> ^e)> value =
Enum.GetValues(typeof<^e>)
|> Seq.cast<^e>
|> Seq.where (fun case -> case &&& value = case)
|> Seq.toList