Develop Reference

Multiple values for single key

I have a list that contains Categories and their accept/reject counts but there is a problem with this list. I'm using a LINQ query to access the data, and I grouped them by both category name and Accept/Reject Code(ResultCode). So the data is in this form:
Almost all of the Categories have both AP counts and RJ counts. And what I'm trying to do is to show each Category's accept and reject count. What should I use? Hashtables don't fit in this problem, I tried Dictionary with int List as value but couldn't add when the same key appeared.
UPDATE:
List<ModReportingDM.ReportObjects.AllCategories> allcats = new List<ModReportingDM.ReportObjects.AllCategories>();
Dictionary<string, ModReportingDM.ReportObjects.ResultCode> dict = new Dictionary<string, ModReportingDM.ReportObjects.ResultCode>();
ModReportingDM.ReportObjects.ResultCode x = new ModReportingDM.ReportObjects.ResultCode();
allcats = reportBLL.GetAllCats(model.ModId, model.ReportStartDate, model.ReportEndDate);
if (allcats != null)
{
model.AllCatsList = new List<ModReportingDM.ReportObjects.AllCategories>();
foreach (var item in allcats)
{
x.Accepted = item.Count;
x.Rejected = item.Count;
dict.Add(item.Category, x);
}
}
Query:
public List<AllCategories> GetAllCats(int modId, DateTime startDate, DateTime endDate)
{
using (entities = new ModReportingEntities())
{
var query = (from c in entities.Content
where c.ModId == modId && c.CreatedTime >= startDate && c.CreatedTime <= endDate && c.Category != null
group c by new { c.Category, c.ResultCode } into g
orderby g.Count() ascending
select new AllCategories
{
Category = g.Key.Category,
ResultCode = g.Key.ResultCode,
AcceptCount = g.Count(),
RejectCount = g.Count()
});
return query.ToList();
}
}

What i would do is create a ResultCode class:
public class ResultCode
{
public int Ap { get; set; }
public int Rj { get; set; }
}
and then use a Dictionary<string, ResultCode> which maps each category to its report.
You could also take a different approach using a Tuple<T1, T2> (which personally i like less) which simply maps your key to two distinct values:
Dictionary<string, Tuple<int, int>> categoryToResultCode;

List<Tuple<string, string, int>> listOfTuples = new List<Tuple<string, string, int>>();
Tuple<string, string, int> tupleItem1 = new Tuple<string, string, int>("A", "AP", 1);
listOfTuples.Add(tupleItem1);
You can use Tuple. Please refer http://msdn.microsoft.com/en-us/library/system.tuple%28v=vs.110%29.aspx

My colleague and I realized that we can keep track of the Category's and if same Category occurs, it means that only a field of it should be changed(either AcceptCount or RejectCount). So we've created a lambda expression like this:
foreach(var item in allcats)
{
if (model.AllCategories.Select(m => m).Where(x => x.Category == item.Category).ToList().Count == 0)
{
if (item.ResultCode == "AP") {
model.AllCategories.Add(new ModReportingDM.ReportObjects.AllCategories()
{
Category = item.Category,
AcceptCount = item.Count
});
}
else
{
model.AllCategories.Add(new ModReportingDM.ReportObjects.AllCategories()
{
Category = item.Category,
RejectCount = item.Count
});
}
}
else
{
ModReportingDM.ReportObjects.AllCategories x = model.AllCategories.Select(n => n).Where(y => y.Category == item.Category).ToList().First();
if (item.ResultCode == "AP")
{
x.AcceptCount = item.Count;
}
else
{
x.RejectCount = item.Count;
}
}
}
If same Category occurs, go ahead and change its AcceptCount or RejectCount accordingly. That's the way I solved the problem.

Linq Conditional .Any() Select

How can I perform a conditional select on a column value, where I have a preference over which value is returned. If I can't find the top choice, I settle on the next, if available, and then if not the next, etc. As it looks right now, it would take 3 total queries. Is there a way to simplify this further?
var myResult = string.Empty;
if (myTable.Where(x => x.ColumnValue == "Three").Any())
{
myResult = "Three"; // Can also be some list.First().Select(x => x.ColumnValue) if that makes it easier;
}
else if (myTable.Where(x => x.ColumnValue == "One").Any())
{
myResult = "One";
}
else if (myTable.Where(x => x.ColumnValue == "Two").Any())
{
myResult = "Two";
}
else
{
myResult = "Four";
}

You could use a string[] for your preferences:
string[] prefs = new[]{ "One", "Two", "Three" };
string myResult = prefs.FirstOrDefault(p => myTable.Any(x => x.ColumnValue == p));
if(myResult == null) myResult = "Four";
Edit Enumerable.Join is a very efficient hash table method, it also needs only one query:
string myResult = prefs.Select((pref, index) => new { pref, index })
.Join(myTable, xPref => xPref.pref, x => x.ColumnValue, (xPref, x) => new { xPref, x })
.OrderBy(x => x.xPref.index)
.Select(x => x.x.ColumnValue)
.DefaultIfEmpty("Four")
.First();
Demo

I wrote an extension method that effectively mirrors Tim Schmelter's answer (was testing this when he posted his update. :-()
public static T PreferredFirst<T>(this IEnumerable<T> data, IEnumerable<T> queryValues, T whenNone)
{
var matched = from d in data
join v in queryValues.Select((value,idx) => new {value, idx}) on d equals v.value
orderby v.idx
select new { d, v.idx };
var found = matched.FirstOrDefault();
return found != null ? found.d : whenNone;
}
// usage:
myResult = myTable.Select(x => x.ColumnValue)
.PreferredFirst(new [] {"Three", "One", "Two"}, "Four");
I've written one that will quit a little more early:
public static T PreferredFirst<T>(this IEnumerable<T> data, IList<T> orderBy, T whenNone)
{
// probably should consider a copy of orderBy if it can vary during runtime
var minIndex = int.MaxValue;
foreach(var d in data)
{
var idx = orderBy.IndexOf(d);
if (idx == 0) return d; // best case; quit now
if (idx > 0 && idx < minIndex) minIndex = idx;
}
// return the best found or "whenNone"
return minIndex == int.MaxValue ? whenNone : orderBy[minIndex];
}

I use a weighted approach in SQL where I assign a weight to each conditional value. The solution would then be found by finding the highest or lowest weight depending on your ordering scheme.
Below would be the equivalent LINQ query. Note that in this example I am assigning a lower weight a higher priority:
void Main()
{
// Assume below list is your dataset
var myList =new List<dynamic>(new []{
new {ColumnKey=1, ColumnValue ="Two"},
new {ColumnKey=2, ColumnValue ="Nine"},
new {ColumnKey=3, ColumnValue ="One"},
new {ColumnKey=4, ColumnValue ="Eight"}});
var result = myList.Select(p => new
{
ColVal = p.ColumnValue,
OrderKey = p.ColumnValue == "Three" ? 1 :
p.ColumnValue == "One" ? 2 :
p.ColumnValue == "Two" ? 3 : 4
}).Where(i=> i.OrderKey != 4)
.OrderBy(i=>i.OrderKey)
.Select(i=> i.ColVal)
.FirstOrDefault();
Console.WriteLine(result ?? "Four");
}

How about something like this:
var results = myTable.GroupBy(x => x.ColumnValue).ToList();
if (results.Contains("Three")) {
myResult = "Three";
} else if (results.Contains("One")) {
myResult = "One";
} else if (results.Contains("Two")) {
myResult = "Two";
} else {
myResult = "Four";
}

Search an Object list for duplicate keys, and place string values into a string array

What I want to be able to do is search the searchValues list for duplicate itemId's, and when I find them, place the single string value into the string values array.
The SearchValue object:
public class SearchValue<TItemId>
{
public TItemId ItemId { get; set; }
public string Value { get; set; }
public string[] Values { get; set; }
}
My test search values after init looks like this:
searchValues[0]
.ItemId == 16
.Value == "2"
searchValues[1]
.ItemId == 16
.Value == "3"
searchValues[2]
.ItemId == 15
.Value == "6"
searchValues[3]
.ItemId == 15
.Value == "3"
searchValues[4]
.ItemId == 5
.Value == "Vertonghen"
I'd like my final result to look like this:
searchValues[0]
.ItemId == 16
.Values == "2,3"
searchValues[1]
.ItemId == 15
.Values == "6,3"
searchValues[2]
.ItemId == 5
.Value == "Vertonghen"
I'd really want to do this using LINQ. I've managed to create another List of SearchValue thusly:
List<SearchValue<Byte>> duplicateSearchItems = (from x in searchValues
group x by x.ItemId into grps
orderby grps.Key
where grps.Count() > 1
select grps).SelectMany(group => group).ToList();
...but getting the value into the values array is giving me trouble. Ideally if the LINQ could return a single list that contains the duplicate records transmogrified into the array with the non-duplicates intact would be best. Probably a nested query of some sort? I'm stumped. Any ideas?
Thanks in advance!

why don't you use Dictionary
Dictionary<int,string> d = new Dictionary<int,string>();
foreach(var x in searchValues)
{
if(d.ContainsKey(x.ItemId))
d[x.ItemId] = string.Format("{0},{1}",d[x.ItemId],x.Value);
else
d.Add(x.ItemId,x.Value);
}
on the end simply iterate throug the Dictionary
foreach(var entry in d)
{
ConsoleWriteline(entry.Key+" : "+entry.Value);
}

Linq alone cannot be used to modify the original list or to modify the items in the list. However, you could do this to create a new list:
List<SearchValue<Byte>> results =
(from x in searchValues
group x by x.ItemId into g
select new SearchValue<Byte>()
{
ItemId = g.Key,
Value = g.Value.First().Value,
Values = g.Value.Select(i => i.Value).ToArray(),
}
.ToList();
Or in fluent syntax:
List<SearchValue<Byte>> results = searchValues
.GroupBy(x => x.ItemId)
.Select(g => new SearchValue<Byte>()
{
ItemId = g.Key,
Value = g.Value.First().Value,
Values = g.Value.Select(i => i.Value).ToArray(),
})
.ToList();
However, depending on your situation, an ILookup may be more appropriate for you:
var results = searchValues.ToLookup(x => x.ItemId, x => x.Value);
Console.Write(String.Join(", ", results[16])); // 2, 16

I don't think LINQ would provide you with the best solution here. Similar to Nikola, I'd use a Dictionary. If you aren't married to your SearchValue data type, you can avoid the second loop that just pushes your data back into your type. A Dictionary> would work here.
var searchValues = new List<SearchValue<int>>();
var distinctItemIds = new Dictionary<int, List<string>>();
foreach (var item in searchValues)
{
if (!distinctItemIds.ContainsKey(item.ItemId))
{
distinctItemIds.Add(item.ItemId, new List<string>());
}
// Add the value
distinctItemIds[item.ItemId].Add(item.Value);
}
// Put values back into your data object
var finishedValues = new List<SearchValue<int>>();
foreach (var keyValuePair in distinctItemIds)
{
finishedValues.Add(new SearchValue<int>()
{
ItemId = keyValuePair.Key,
Values = keyValuePair.Value.ToArray()
});
}

I managed to work it out using LINQ.
// Get a new list of unique items to add our duplicated items to
List<SearchValue<Byte>> finalSearchItems = (from x in searchValues
group x by x.ItemId into grps
orderby grps.Key
where grps.Count() == 1
select grps).SelectMany(group => group).ToList();
byte[] duplicateIds = (from x in searchValues
group x by x.ItemId into grps
where grps.Count() > 1
select grps.Key).ToArray();
// Smash the string 'Value' into 'Values[]'
foreach (byte id in duplicateIds)
{
SearchValue<Byte> val = new SearchValue<byte>();
val.ItemId = id;
// Smash
val.Values = (from s in searchValues
where s.ItemId == id
select s.Value).ToArray();
finalSearchItems.Add(val);
}

Remove from List<> duplicated items and record duplicated quantity

I have a List<Item>. Item has properties Id,Name and Amount. There are duplicated items in this list. I need to get a new List<Item> which contains only non-duplicated Items and in Item's Amount should be the quantity of how many times it duplicated in first List<Item>. I tried something like
for (int i = 0; i < list.Count; i++)
{
for (int j = 0; j < list.Count; j++)
{
if (list[i].Name == list[j].Name)
{
list.Remove(prod.Components[j]);
list[i].Amount++;
}
}
}
but there are some problems in this loop. My brains are overheated. Please, help.

A simple LINQ query can get you the unique items along with the number of times they appear:
var distinct = list.GroupBy(o => o.Name)
.Select(g => new { Count = g.Count(), Item = g.First() })
.ToList();
Then you can modify each item's Amount to the count of duplicates:
foreach (var row in distinct)
{
row.Item.Amount = row.Count;
}
And finally get back a List<Item> that contains no duplicates and has the correct amounts:
var uniqueItems = distinct.Select(r => r.Item).ToList();
Important: The code above assumes that "duplicate" items are indistinguishable from each other, but nothing else (e.g. it doesn't need Item to have a default constructor). Depending on the particulars it may be possible to write it in an even shorter form.
Additionally, the Amount property looks strange here. Since duplicates do not warrant summation of their amounts, what's the purpose of Item.Amount? I would assume that duplicate items with amount of 2 should result in one item with an amount of 4, but your code does not do that (and mine follows that lead).

Off the top of my head (haven't tested it):
list.GroupBy(x => x.Name)
.Select(x => new Item {
Name = x.Key,
Amount = x.Count()
})
.ToList();
You haven't specified what happens to the Ids, so I've left ignored them.
(Note this creates a new list, rather than modifying the original).

Try something like that:
var groups = from item in items
group item by item.Property
into grouped
select grouped;
var distinct = from g in groups
let item = g.First()
let amount = g.Count()
select new Item {Property = item.Property, Amount = amount};
After that distinct contains IEnumerable<Item> with their amount from original items list.

foreach(var item in list)
{
if(list.Count(e=>e.Id == item.Id && e.Name == item.Name)!=1)
{
list.Remove(item);
}
}

Assuming that you determine duplicates by the first two properties ID and Name.
You can implement an IEqualityComparer<Item> and use that for Enumerable.GroupBy:
var itemAmounts = items.GroupBy(i => i, new Item())
.Select(g => new Item {
ID = g.First().ID,
Name = g.First().Name,
Amount = g.Count()
});
Here's your Item class with a meaningful implementation of IEqualityComparer<Item>:
class Item : IEqualityComparer<Item>
{
public int ID;
public string Name;
public int Amount;
public bool Equals(Item x, Item y)
{
if (x == null || y == null) return false;
bool equals = x.ID == y.ID && x.Name == y.Name;
return equals;
}
public int GetHashCode(Item obj)
{
if (obj == null) return int.MinValue;
int hash = 19;
hash = hash + obj.ID.GetHashCode();
hash = hash + obj.Name.GetHashCode();
return hash;
}
}
You could also override Equals and GetHasdhCode from object, then you don't need a custom comparer at all in GroupBy:
var itemAmounts = items.GroupBy(i => i)
.Select(g => new Item {
ID = g.First().ID,
Name = g.First().Name,
Amount = g.Count()
});
You can use above already available methods:
public override bool Equals(object obj)
{
Item item2 = obj as Item;
if (item2 == null)
return false;
else
return Equals(this, item2);
}
public override int GetHashCode()
{
return GetHashCode(this);
}

Sort list by field (C#)

I have class like:
class SortNode
{
public Int32 m_valRating = 0;
public SortNode(Int32 valRating)
{
this.m_valRating = valRating;
}
}
and some list refSortNodeList:
List<SortNode> refSortNodeList = new List<SortNode>();
Random refRandom = new Random();
for (int i = 0; i < 100; ++i)
{
refSortNodeList.Add(new SortNode(refRandom.Next(-10, 30)));
}
foreach (var varSortNode in refSortNodeList)
{
Console.WriteLine("SortNode rating is {0}", varSortNode.m_valRating);
}
How to sort easily my refSortNodeList by m_valRating field? Or maybe I need to use some another List class?

list.Sort((x,y) =>
x.m_valRating.CompareTo(y.m_valRating));

In-place:
refSortNodeList.Sort(
(x, y) =>
x == null ? (y == null ? 0 : -1)
: (y == null ? 1 : x.m_valRating.CompareTo(y.m_valRating))
);
Creating a new enumeration:
var newEnum = refSortNodeList.OrderBy(x => x.m_valRating);
Creating a new list:
var newList = refSortNodeList.OrderBy(x => x.m_valRating).ToList();
In-place is fastest and most memory efficient, but no good if you want to also retain the old list.
The next is faster than the last and gives results as they go, but you have to re-do the sort to use it again, in which case the third is the one to go for.

Use Linq order by.
var mySortedList = refSortNodeList.OrderBy(x => x.m_valRating);
Here is a real live example where I am pulling a list from a database but it is exactly the same concept.
vendorProducts = (from vp in db.COMPANIES_VND_PRODUCTS
join p in db.CT_CT_INV_CLASSES on vp.CLASS_ID equals p.CLASS_ID
join m in db.CT_CT_MODALITY_CODES on vp.MODALITY_ID equals m.MODALITY_ID
where vp.COMPANY_ID == companyId
select new ProductTypeModality
{
Active = p.ACTIVE.Equals("Y") ? true : false,
BioMedImaging = p.BIOMED_IMAGING,
Code = p.CLASS_CODE,
Description = p.DESCRIPTION,
Id = p.CLASS_ID,
PricingMargin = p.PRICING_MARGIN,
ModalityCode = m.MODALITY_CODE,
ModalityId = m.MODALITY_ID,
VendorId = companyId
}).OrderBy(x => x.Code).ToList<ProductTypeModality>();

Implement IComparable<T>

You can use Linq for basic sorts:
refSortNodeList.OrderBy(n => n.m_valRating);
If you need more complex sorting your will need to implement IComparable to use the built in sorting.

Try this:
refSortNodeList.Sort(new delgate(SortNode x, SortNode y)
{
return x.CompareTo(y);
}
);

It's easy using linq:
var newlist = refSortNodeList.sort( n => n.m_valRating );

List<SortNode> refSortNodeList = new List<SortNode> ();
Random refRandom = new Random ();
for (int i = 0; i < 100; ++i) {
refSortNodeList.Add (new SortNode (refRandom.Next (-10, 30)));
}
// Use this (Linq) if you're using .NET 3.5 or above.
var sortedList = refSortNodeList.OrderBy (node => node.m_valRating);
foreach (var varSortNode in sortedList) {
Console.WriteLine ("SortNode rating is {0}", varSortNode.m_valRating);
}
// Use this otherwise (e.g. .NET 2.0)
refSortNodeList.Sort (
delegate (SortNode n1, SortNode n2) {
return n1.m_valRating.CompareTo (n2.m_valRating);
}
);
foreach (var varSortNode in refSortNodeList) {
Console.WriteLine ("SortNode rating is {0}", varSortNode.m_valRating);
}