DataTable Query - linq

I am new to LINQ. I am trying to find the rows that does not exists in the second data table.
report_list and benchmark both type are : DataTable. Both these datatables are being populated using OleDbCommand,OleDbDataAdapter. I am getting an error "Specified cast is not valid." in foreach ... loop. I would appreciate your help.
var result = from a in report_list.AsEnumerable()
where !(from b in benchmark.AsEnumerable()
select b.Field<int>("bench_id")
)
.Contains(a.Field<int>("BenchmarkID"))
select a;
foreach (var c in result)
{
Console.WriteLine(c.Field<string>("Name"));
}

I don't know if I understood your question. Are you trying to get the items that exists in the first table but not in the second?
var first = new string[] { "b", "c" };
var second = new string[] { "a", "c" };
//find the itens that exist in "first" but not in "second"
var q = from f in first
where !second.Contains(f)
select f;
foreach (var s in q) {
Console.WriteLine(s);
}
//Prints:
//b
I suggest you to make the inner query first, once it does not depend on the outer record.

From a in report_list
Group Join b in benchmark On a.bench_id Equals b.bench_id Into g = Group
Where g.Count = 0
Select a
Note that this is VB syntax.

My suspicion is that one of the fields you are comparing is not an integer in the database. I believe that the invalid cast exception is being thrown by one of the Field<int>() calls since that is one of the three different exceptions that this method can throw. See docs here.

Perhaps use the .Except() extension to get the set difference of the two sets?
(from b in benchmark.AsEnumerable()
select new { id = b.Field<int>("bench_id")}).Except(
from a in report_list.AsEnumerable()
select new {id = a.Field<int>("BenchmarkID")})
Not actually sure of the precise syntax, but that should work by taking the ids in benchmark, and then removing all equivalent ids in report_list, leaving only the ids that don't match. (I hope this is the order you were after...)
Note: This is also assuming that the above issue mentioned by tvanfosson isn't also a problem

Related

Combining LINQ Queries to reduce database calls

I have 2 queries that work, I was hoping to combine them to reduce the database calls.
var locations = from l in db.Locations
where l.LocationID.Equals(TagID)
select l;
I do the above because I need l.Name, but is there a way to take the above results and put them into the query below?
articles = from a in db.Articles
where
(
from l in a.Locations
where l.LocationID.Equals(TagID)
select l
).Any()
select a;
Will I actually be reducing any database calls here?
This seems a bit complicated because Locations appears to be a multi-value property of Articles and you want to only load the correct one. According to this answer to a similar question you need to use a select to return them separately in one go so e.g.
var articles = from a in db.Articles
select new {
Article = a,
Location = a.Locations.Where(l => l.LocationId == TagId)
};
First failed attempt using join:
var articlesAndLocations = from a in db.Articles
join l in a.Locations
on l.LocationID equals TagID
select new { Article = a, Location = l };
(I usually use the other LINQ syntax though so apologies if I've done something stupid there.)
Could you not use the Include() method here to pull in the locations which are associated with each article, then select both the article and location object? or the properties you need from each.
The include method will ensure that you don't need to dip into the db twice, but will allow you to access properties on related entities.
You would need to use a contains method on an IEnumerable I believe, something like this:
var tagIdList = new List() { TagID };
var articles = from a in db.Articles.Include("Locations")
where tagIdList.Contains(from l in a.Locations select l.LocationID)
select new { a, a.Locations.Name };
(Untested)

retrieve all areas where id's don't exist in supplied list

I'm sure I must be missing something really simple here..
OK I have a list of AreaIds. I want to compare that list to the MapArea Table and return any IDs that exist in the table but NOT in the supplied list.
This is my list of supplied areas that I want to check:
var currentAreas = (from c in _entities.mapAreaLink
where c.listingId == id
select new
{
c.MapArea.areaId
}
).ToList();
This is the getting the exhaustive list of mapAreas..
var availableAreas = (from m in _entities.MapAreas
select new
{
m.areaId
}
).ToList();
This compares the two lists and gets items that exist in the maparea table but not in the maparealink (constrained by an id of the item I am looking at).
var unusedAreas = availableAreas.Except(currentAreas).ToList();
I seem to get the list back ok, but what I need to do is now return a list of maparea objects based on the results of the Except.tolist above.
I thought I could do this:
var mapareas = (from e in _entities.MapAreas
where unusedAreas.Contains(e.areaId)
select e).ToList();
I am getting an ambiguous invocation on the where & "Cannot resolve method Contains(int)" on the e.areaId.
Ive tried using:
var unusedAreas = availableAreas.Except(currentAreas).ToArray();
No Joy.. Can anyone help me out here - I am guessing I must be missing a fundamental basic here.
many thanks
You create anonymous types with just one int property. That's not necessary and it causes the later problem. If you create lists of int you'll be OK:
var currentAreas = (from c in _entities.mapAreaLink
where c.listingId == id
select c.MapArea.areaId).ToList();
var availableAreas = (from m in _entities.MapAreas
select m.areaId).ToList();

LINQ subquery question

Can anybody tell me how I would get the records in the first statement that are not in the second statement (see below)?
from or in TblOrganisations
where or.OrgType == 2
select or.PkOrgID
Second query:
from o in TblOrganisations
join m in LuMetricSites
on o.PkOrgID equals m.FkSiteID
orderby m.SiteOrder
select o.PkOrgID
If you only need the IDs then Except should do the trick:
var inFirstButNotInSecond = first.Except(second);
Note that Except treats the two sequences as sets. This means that any duplicate elements in first won't be included in the results. I suspect that this won't be a problem since the name PkOrgID suggests a unique ID of some kind.
(See the documentation for Enumerable.Except and Queryable.Except for more info.)
Do you need the whole records, or just the IDs? The IDs are easy...
var ids = firstQuery.Except(secondQuery);
EDIT: Okay, if you can't do that, you'll need something like:
var secondQuery = ...; // As you've already got it
var query = from or in TblOrganisations
where or.OrgType == 2
where !secondQuery.Contains(or.PkOrgID)
select ...;
Check the SQL it produces, but I think it should do the right thing. Note that there's no point in performing any ordering in the second query - or even the join against TblOrganisations. In other words, you could use:
var query = from or in TblOrganisations
where or.OrgType == 2
where !LuMetricSites.Select(m => m.FkSiteID).Contains(or.PkOrgID)
select ...;
Use Except:
var filtered = first.Except(second);

List.GroupBy<> error using LInq

I'm trying to group a generic List<> in C#. The code compiles, but the application (Silverlight) throws the following error (CharOpps is the class of objects in the list I'm trying to group):
Unhandled Error in Silverlight Application Unable to cast object of type 'Grouping[System.DateTime,Invoc_SalesDashboard.ChartOpps]' to type 'Invoc_SalesDashboard.ChartOpps'.
Here's the code:
var newtemplist = list.GroupBy(opp =>
new DateTime(opp.EstimatedCloseDate.Year, opp.EstimatedCloseDate.Month, 1)).OrderBy(opp => opp.Key);
I've also tried:
var newtemplist =
from opp in list
orderby opp.EstimatedCloseDate
group opp by new { opp.EstimatedCloseYear, opp.EstimatedCloseMonth };
ChartOpps have a revenue value, and the EstimatedCloseDate value. What I'm hoping to end up with is a list of ChartOpps with the revenue aggregated in the year/month groupings.
foreach (ChartOpps c in newtemplist)
{
ErrorBox.Text += "o";
}
You're not showing us what you're doing with the result newtemplist. The runtime error message indicates that you are taking a group and trying to treat it as an instance of ChartOpps which is clearly impossible. Show us that code and we can help you fix it.
Edit:
Okay, now the problem is clear. To enumerate the results of the grouping, you need to proceed as follows:
foreach(var group in newtemplist) {
foreach(ChartOpps c in group) {
// do something with c here
}
}
The result of newtemplist is a sequence of sequences, each sequence having all of its elements having the same value of new DateTime(opp.EstimatedCloseDate.Year, opp.EstimatedCloseDate.Month, 1). Therefore, to enumerate this sequence of sequences, you first have to enumerate the groups, and then within each group enumerate the instances of ChartOpps.
Not knowing anything about your class structure, here is a basic attempt:
List<CharOpps> list = GetList();
var newtemplist =
from opp in list
group opp by opp.EstimatedCloseYear into g
select new { g = g.Key, CharOpps = g };
If you take the var out of the picture, it all becomes clear.
IEnumerable<IGrouping<DateTime, ChartOpps>> newtemplist = list
.GroupBy(opp => new DateTime(
opp.EstimatedCloseDate.Year,
opp.EstimatedCloseDate.Month,
1))
.OrderBy(opp => opp.Key);
foreach (ChartOpps c in newtemplist)
{
ErrorBox.Text += "o";
}
The error occurs in the assignment of the first element of newtemplist to c. c is allowed to reference ChartOpps instances. The first element of newtemplist is a IGrouping<DateTime, ChartOpps>, not a ChartOpps. The implicit cast in the foreach fails and you get a runtime exception.
Try instead:
foreach(IGrouping<DateTime, ChartOpps> g in newtemplist)
{
foreach (ChartOpps c in g)
{
ErrorBox.Text += "o";
}
}

Linq to SQL: .FirstOrDefault() not applicable to select new { ... }

I just asked this question. Which lead me to a new question :)
Up until this point, I have used the following pattern of selecting stuff with Linq to SQL, with the purpose of being able to handle 0 "rows" returned by the query:
var person = (from p in [DataContextObject].Persons
where p.PersonsID == 1
select new p).FirstOrDefault();
if (person == null)
{
// handle 0 "rows" returned.
}
But I can't use FirstOrDefault() when I do:
var person = from p in [DataContextObject].Persons
where p.PersonsID == 1
select new { p.PersonsID, p.PersonsAdress, p.PersonsZipcode };
// Under the hood, this pattern generates a query which selects specific
// columns which will be faster than selecting all columns as the above
// snippet of code does. This results in a performance-boost on large tables.
How do I check for 0 "rows" returned by the query, using the second pattern?
UPDATE:
I think my build fails because I am trying to assign the result of the query to a variable (this._user) declared with the type of [DataContext].User.
this._user = (from u in [DataContextObject].Users
where u.UsersID == [Int32]
select new { u.UsersID }).FirstOrDefault();
Compilation error: Cannot implicitly convert type "AnonymousType#1" to "[DataContext].User".
Any thoughts on how I can get around this? Would I have to make my own object?
Why can you keep doing the samething? Is it giving you an error?
var person = (from p in [DataContextObject].Persons
where p.PersonsID == 1
select new { p.PersonsID, p.PersonsAdress, p.PersonsZipcode }).FirstOrDefault();
if (person == null) {
// handle 0 "rows" returned.
}
It is still a reference object just like you actual object, it is just anonymous so you don't know the actual type before the code is compiled.
Update:
I see now what you were actually asking! Sorry, my answer no longer applies. I thought you were not getting a null value when it was empty. The accepted response is correct, if you want to use the object out of scope, you need to create a new type and just use New MyType(...). I know DevEx's RefactorPro has a refactoring for this, and I think resharper does as well.
Call .FirstOrDefault(null) like this:
string[] names = { "jim", "jane", "joe", "john", "jeremy", "jebus" };
var person = (
from p in names where p.StartsWith("notpresent") select
new { Name=p, FirstLetter=p.Substring(0,1) }
)
.DefaultIfEmpty(null)
.FirstOrDefault();
MessageBox.Show(person==null?"person was null":person.Name + "/" + person.FirstLetter);
That does the trick for me.
Regarding your UPDATE: you have to either create your own type, change this._user to be int, or select the whole object, not only specific columns.
if (person.Any()) /* ... */;
OR
if (person.Count() == 0) /* ... */;
You can still use FirstOrDefault. Just have
var PersonFields = (...).FirstOrDefault()
PersonFields will be be null or an object with those properties you created.

Resources