Two ways:
- Methods which return expressions can be used
- Separate the queryable and enumerable bits
For #1, consider:
public Expression<Func<Foo, bool>> WhereCreatorIsAdministrator()
{
return f => f.Creator.UserName.Equals("Administrator", StringComparison.OrdinalIgnoreCase);
}
public void DoStuff()
{
var exp = WhereCreatorIsAdministrator();
using (var c = new MyEntities())
{
var q = c.Foos.Where(exp); // supported in L2E
// do stuff
}
}
For an example of number 2, read this article: How to compose L2O and L2E queries. Consider the example given there:
var partialFilter = from p in ctx.People
where p.Address.City == “Sammamish”
select p;
var possibleBuyers = from p in partiallyFilter.AsEnumerable()
where InMarketForAHouse(p);
select p;
This might be less efficient, or it might be fine. It depends on what you're doing. It's usually fine for projections, often not OK for restrictions.
Update Just saw an even better explanation of option #1 from Damien Guard.
与恶龙缠斗过久,自身亦成为恶龙;凝视深渊过久,深渊将回以凝视…