Question

I am simulating a join on a linked server through linq to sql. My question is, it appears that linqtosql is bringing all the rows for y.Xstatuses into memory and then doing the join. If this is true how do i keep all the memory on sql server(and still do a cross datacontext join operation), if this not true what is going on that is eating all my ram?

var x = new fooDataContext();
var y = new barDataContext();

var allXNotDeleted = (from w in x.CoolTable
                      where x.IsDeleted != false).ToList();//for our demo this returns 218 records

var allXWithCompleteStatus = (from notDeleted in allXNotDeleted
                              join s in y.XStatuses on notDeleted.StatusID equals s.StatusID
                              where s.StatusID == 1
                              select notDeleted).Tolist();// insert massive memory gobbler here

return allXwithCompleteStatus;

EDIT:
Trying to implement Kevinbabcock’s idea

    using (x = new fooDataContext())
    using (var y = new barDataContext())
    {

        var n = (from notDeleted in x.GetTable<CoolTable>()
             join z in y.GetTable<Xstatus>() on x.StatusID equals z.StatusID
             where z.StatusID == 1 and x.IsDeleted != false
             select x).ToList();

    }

This still throws a cross context query exeception

Answer 1

It is not possible to perform cross data context query directly on the database.
Fetching in memory one of the recordset (ToList()) forces anyway the other joined to be processed in memory.

If you want to perform everything on sql server you have to have every entity in the same DataContext.