Hi I\'m using linq to entity in my application. I need to get distinct records based on one column value \"Name\"
So I have a table similar like you can see below:
Hi here is how you can select distinct records with inner join. Hope it helps
var distinctrecords =
(entity.Table.Join(entity.Table2, x => x.Column, y => y.Column, (x, y) => new {x, y})
.Select(@t => new {@t.x.Column2, @t.y.Column3}))
.GroupBy(t => t.Column2)
.Select(g => g.FirstOrDefault());
You can use something like this:
var distinctReports = reports.Select(c => c.CompanyCode)
.Distinct()
.Select(c => reports.FirstOrDefault(r => r.CompanyCode == c))
.ToList();
The Distinct()
method doesn't perform well because it doesn't send the DISTINCT SQL predicate to the database. Use group
instead:
var distinctResult = from c in result
group c by c.Id into uniqueIds
select uniqueIds.FirstOrDefault();
LINQ's group actually creates subgroups of entities keyed by the property you indicate:
Smith
John
Mary
Ed
Jones
Jerry
Bob
Sally
The syntax above returns the keys, resulting in a distinct list. More information here:
http://imar.spaanjaars.com/546/using-grouping-instead-of-distinct-in-entity-framework-to-optimize-performance
The purely LINQ way that occurs is to group by name, select distinct groups by key, then select based on that.
from i in user
group new {i.ID, i.Country, i.DateRecord} by i.Name into byNmGp
select byNmGp.First();
Edit: Entity Framework is of course a very popular linq provider, but it doesn't handle First()
well here, though the logically equivalent (in this case) FirstOrDefault()
will work fine. I prefer First()
when not forced into FirstOrDefault()
by EF's limitations, because its meaning better matches what is sought here.
Another way is to define a helper class:
private class MyRecord : IEquatable<MyRecord>
{
public int ID;
public string Name;
public string Country;
public DateTime DateCreated;
public bool Equals(MyRecord other)
{
return Name.Equals(other.Name);
}
public override bool Equals(object obj)
{
return obj is MyRecord && Equals((MyRecord)obj);
}
public override int GetHashCode()
{
return Name.GetHashCode();
}
}
/*...*/
var items = (from i in user select new MyRecord {i.ID, i.Name, i.Country, i.DateRecord}).Distinct();
This simply defines distinct differently. Performance will differ by whether the query provider can interpret that definition of equality or not. Convenience will differ based on whether you've similar LINQ queries doing much the same thing or not.
Here's another variation I ended up using which was based off the response from Svetlana. Shows an example of populating a GridView control with unique values. Thanks!
dataGridView_AnalyzeTestSuites.DataSource = (
from tr in _db.TestResults
where tr.TaskId == taskId
select new { TestSuiteName = tr.Test.TestSuite.Name }
).Distinct().ToList();