Earlier I put a question on Stackoverflow about how to remove duplicate records in a list of objects, based on a particular property within each object.
I got the answer I was looking for (see below), a query which returns a distinct list of objects using MainHeadingID as the property to remove duplicates.
public IList<tblcours> GetAllCoursesByOrgID(int id)
{
return _UoW.tblcoursRepo.All.
Where(c => c.tblCourseCategoryLinks.Any(cl => cl.tblUnitCategory.tblUnit.ParentID == id))
.GroupBy(c => c.MainHeadingID)
.Select(g => g.FirstOrDefault())
.ToList();
}
However, now I need more help! Is there anyway of amending the query above so that, it only removes duplicate values when MainHeadingID is not equal to 180. I tried amending GroupBy line to
.GroupBy(c => c.MainHeadingID != 180)
However, this didn't work.
Any help would be much appreciated with this.
Thanks.
Following works for LINQ to SQL:
return _UoW.tblcoursRepo.All
.Where(c => c.tblCourseCategoryLinks.Any(cl => cl.tblUnitCategory.tblUnit.ParentID == id))
.GroupBy(c => c.MainHeadingID)
//.SelectMany(g => g.Key == 180 ? g : g.Take(1))
.SelectMany(g => g.Take(g.Key == 180 ? Int32.MaxValue : 1))
.ToList();
Comments: SelectMany in query above selects all items from group where MainHeadingID equals to 180, but it takes only one item form other groups (i.e. distinct result). Linq to SQL cannot translate commented out part, but thanks to #usr there is way around.
Linq to Entities cannot translate even simplified query. I think only option for you in this case is simple concating result of two queries:
Expression<Func<tblcours, bool>> predicate = x =>
x.tblCourseCategoryLinks.Any(cl => cl.tblUnitCategory.tblUnit.ParentID == id)
int headingId = 180;
return _UoW.tblcoursRepo.All
.Where(c => c.MainHeadingID != headingId)
.Where(predicate)
.GroupBy(c => c.MainHeadingID)
.Select(g => g.FirstOrDefault())
.Concat(_UoW.tblcoursRepo.All
.Where(c => c.MainHeadingID == headingId)
.Where(predicate))
.ToList();
lazyberezovsky's answer fails due to an EF bug (which is not surprising given the quality of EF's LINQ support). It can be made to work with a hack:
.SelectMany(g => g.Key == 180 ? g.Take(int.MaxValue) : g.Take(1))
or
.SelectMany(g => g.Take(g.Key == 180 ? int.MaxValue : 1))
Note that performance will not be particularly good due to the way this is translated to SQL.
Related
I have a Linq expression that is used in a few places. I went down the expression route as there wasn't a logical way to accomplish some searching logic without enumerating a very large table otherwise.
private Expression<Func<Property, bool>> PropertyIsCompliant()
{
return (p) => p.CalculationSets.OfType<SingleDocumentCalculationSet>()
.GroupBy(cs => cs.SourceDocument)
.Select(g => g.OrderByDescending(d => d.DateTime).FirstOrDefault().CalculationResults)
.SelectMany(cr => cr)
.All(cr => cr.Outcome == CalculationOutcome.Success);
}
My models are as such:
A Property has many CalculationSets
Each CalculationSet is also assigned to a Document
Each CalculationSet has a number of CalculationResults
Each CalculationResult has an Outcome
I'm trying to create an expression that will tell me if all the outcomes from the most recent calculationsets grouped by document ordered by most recent (ie the most recent distinct results) are Successful.
I can the SelectMany clause returns all the CalculationResults from the correct CalculationSets.
I just cant figure out how to return true ONLY if the collection isn't empty AND they are all Outcome.Success.
I understand the All operator automatically returns true on an empty collection. I just can't think of a way around it!
So your real condition is that there are not any unsuccessful outcomes. In that case use Any and reverse the condition:
//V-- notice the ! inverse operator here
return (p) => !(p.CalculationSets.OfType<SingleDocumentCalculationSet>()
.GroupBy(cs => cs.SourceDocument)
.Select(g => g.OrderByDescending(d => d.DateTime).FirstOrDefault().CalculationResults)
.SelectMany(cr => cr)
.Any(cr => cr.Outcome != CalculationOutcome.Success));
var countsBySuccess =
...
.GroupBy(cr => cr.Outcome == CalculationOutcome.Success) //group on success
.Select(g => new { IsSuccessful = g.Key, Count = g.Count() });
You can now examine the two result rows to make sure that the unsuccessful count is zero and the successful count is non-zero.
Regarding performance, this will need to materialize the entire result set server-side and aggregate it. But it does so only once.
If you must use the calculation result as part of a bigger query, you must use another trick:
!countsBySuccess.Any(g =>
g.IsSuccessful && Count == 0 ||
!g.IsSuccessful && Count != 0)
This boolean expression determines whether the condition you are looking for holds with one scan of the data.
It is important to only scan the data once. Do not simply write:
myItems.All(cr => cr.Outcome == CalculationOutcome.Success) && myItems.Any()
Because that does two scans. SQL Server does not optimize this out.
I think you're answering your question - if you know that All returns TRUE for empty then you have two checks to make. Excuse my C# (I'm not sure on the var query assignment, hopefully you get the idea) but you could do something like this:
private Expression<Func<Property, bool>> PropertyIsCompliant()
{
var query = (p) => p.CalculationSets.OfType<SingleDocumentCalculationSet>()
.GroupBy(cs => cs.SourceDocument)
.Select(g => g.OrderByDescending(d => d.DateTime).FirstOrDefault().CalculationResults)
.SelectMany(cr => cr);
return (query.Count > 0) & query.All(cr => cr.Outcome == CalculationOutcome.Success);
}
I didn't realise it was possible to use "&&" in expressions. So I've managed to combine 2 separate expressions that give the answer I need. The "&&" only returns true when both expressions evaluate "true"
return (p) =>
p.CalculationSets.OfType<SingleDocumentCalculationSet>()
.GroupBy(cs => cs.SourceDocument)
.Select(g => g.OrderByDescending(d => d.DateTime).FirstOrDefault().CalculationResults)
.SelectMany(cr => cr).Any()
&&
p.CalculationSets.OfType<SingleDocumentCalculationSet>()
.GroupBy(cs => cs.SourceDocument)
.Select(g => g.OrderByDescending(d => d.DateTime).FirstOrDefault().CalculationResults)
.SelectMany(cr => cr)
.All(cr => cr.Outcome == CalculationOutcome.Success);
I have an xml document
<NumSet>
<num>1</num>
<num>2</num>
<num>2</num>
<num>3</num>
</NumSet>
I want unique elements shown up, ie 1 and 3. not distinct which will also bring out 2.
How to do that? Do I have to use Group? Is there any concise way to do that?
You are right, you can use GroupBy and filter group which has only one item by using Count() == 1:
var output = XDocument.Load(xmlFile)
.Descendants("num")
.Select(e => e.Value)
.GroupBy(x => x)
.Where(g => g.Count() == 1)
.Select(g => g.Key);
It sounds like you want a Distinct GroupBy query... Take a look at the Need help on Linq with group by and distinct post here on StackOverflow.
XElement xe = XElement.Parse(#"<NumSet><num>1</num><num>2</num><num>2</num><num>3</num></NumSet>");
var query = xe.Elements("num")
.GroupBy(x => x.Value)
.Where(x=>x.Count ()==1)
.Select (x => x);
To do what you need I'd say that yes, you need to use GrouBy, and then count the elements in each group, and return those that contains just one element. In code, this translates to:
var query = lst.GroupBy(x => x)
.Where(x => x.Count() == 1)
.Select(x => x.Key);
I am using "Linq" to filter list of objects and to sort them, like
myList.Where(x => x.Item!= "SF" && x.AdSize == minadSize)
.OrderBy(x => x.ManufacturingDate)
.OrderBy(x=>x.ExpiryDate);
I doubt whether i am doing it right or not that is if i want to "sorting" on multiple fields then is it necessary to use multiple Order By clause cant it be done with single "OrderBy"
Don't use multiple OrderBy calls - use OrderBy followed by ThenBy:
var query = myList.Where(x => x.Item!= "SF" && x.AdSize == minadSize)
.OrderBy(x => x.ManufacturingDate)
.ThenBy(x => x.ExpiryDate); // Could add more ThenBy calls
If you use OrderBy twice, it will reorder the already-ordered-by-date list by expiry-date, whereas I assume you only want to order by expiry date for items with an equal manufacturing date, which is what the above does.
Obviously there's a ThenByDescending method too. For example:
var query = people.OrderBy(x => x.LastName)
.ThenBy(x => x.FirstName)
.ThenByDescending(x => x.Age)
.ThenBy(x => x.SocialSecurity);
Linq and EF4.
I have this Linq query in query syntax I would like convert into query method.
Are you able to do it? I tried more tha 2 hours without success :-(
Thanks for your time
CmsContent myContentObj = (from cnt in context.CmsContents
from categoy in cnt.CmsCategories
where categoy.CategoryId == myCurrentCategoryId && cnt.ContentId == myCurrentContentId
select cnt).Single();
My original answer selected the wrong item. It's a bit more complicated than what I had (which Ani has posted). Here's what I believe is an equivalent query however and should perform better:
CmsContent myContentObj =
context.CmsContents
.Where(cnt => cnt.ContentId == myCurrentId
&& cnt.CmsCategories
.Any(categoy => categoy.CategoryId == myCurrentCategoryId))
.Single();
Here is a non-direct translation that I believe performs the same task in much less code:
var myContentObj = context.CmsContents.Single(
x => x.ContentId == myCurrentContentId &&
x.CmsCategories.Any(y => y.CategoryId == myCurrentCategoryId)
);
Here's how the C# compiler actually does it, with some help from .NET Reflector to verify:
var myContentObj = context
.CmsContents
.SelectMany(cnt => cnt.CmsCategories,
(cnt, categoy) => new { cnt, categoy })
.Where(a => a.categoy.CategoryId == myCurrentCategoryId
&& a.cnt.ContentId == myCurrentContentId)
.Select(a => a.cnt)
.Single();
Essentially, the 'nested' from clauses results in a SelectMany call with a transparent identifier (an anonymous-type instance holding the 'parent' cnt and the 'child' categoy). The Where filter is applied on the anonymous-type instance, and then we do another Select projection to get back the 'parent'. The Single call was always 'outside' the query expression of course, so it should be obvious how that fits in.
For more information, I suggest reading Jon Skeet's article How query expressions work.
I have a linq expression that returns transactions in groups. Each transaction has a numerical value and I now need to know what is the highest value from all the transactions returned. This value is held in a field called TransactionId
Here is the expression I am using to get the grouped list.
var transactions = ctx.MyTransactions
.Where (x => x.AdapterId == Id)
.GroupBy(x => x.DeviceTypeId);
I now need to write an expression that works on the “transactions” grouped list to find the “max” of the TransactionId field. I’ve tried different ideas but none seem to work with the grouped results. I’m new to linq so I’m not sure how to do this.
Have you tried finding the maximum in each group and then finding the maximum of that over all groups?
int max = transactions.Max(g => g.Max(t => t.TransactionId));
Or you could just query the database again:
int max = ctx.MyTransactions
.Where(x => x.AdapterId == Id)
.Max(t => t.TransactionId);
This will give you the max in each group
var transactionIds = ctx.MyTransactions
.Where (x => x.AdapterId == Id)
.GroupBy(x => x.DeviceTypeId,
g => new {
DeviceTypeId = g.Key,
MaxTransaction = g.Max(x => x.TransactionId)
});