Linq查询嵌套选择+不同

Question

I'm having a really hard time converting this query to LINQ.

select fk, count(*)
from (
    select distinct fk, attribute1, attribute2
    from table
) a
group by fk
having count(*) > X

I need each distinct combination of fk, attr1, attr2 grouped by fk only where count is bigger than x (parameter).

Any ideas?

Answer 1

First, create your nested query:

var query= db.Table.Select(e=>new {e.FK, e.Attribute1, e.Attribute2}).Distinct();

After that, you can do this:

int x=10;
var result= from e in query
            group e by e.FK into g
            let count=g.Count()
            where count>x
            select new{FK=g.Key, Count=count};

When you call, for example, the ToList method, is when those queries are going to be executed against your DB and the resulted elements are going to be loaded into the memory.

var elements=result.ToList();

Merging all, you could also do this:

int x=10;
var elements=(from e in db.Table.Select(e=>new {e.FK, e.Attribute1, e.Attribute2}).Distinct()
             group e by e.FK into g
             let count=g.Count()
             where count>x
             select new{FK=g.Key, Count=count}).ToList();

Answer 2

You'll have to implement IComparable , IComparer or pass IEqualityComparer or some other way to ensure Distinct() works across your class' properties, but here's the essential linq:

var result = (from t in tableList.Select(t => new { fk, attr1, attr2 }).Distinct()
              group t by t.fk into g
              select new { fk = g.Key, Count = g.Count() })
             .Where(t => t.Count > X)
             .ToList();

this uses an anonymous type as the equivalent of 'select fk, attribute1, attribute2' but that may make it harder to implement Distinct().

By default Distinct() does an object-level compare (are the two instances the same instance) rather than a value-level compare (do they have the same value).