Linq: Skip(0) & Take(0)

[LeaFrock]

I notice that Queryable will still continue to create new query when Skip(0)/Take(0) is called.
Is that necessary? In my opinion, it should be optimized.

[huoyaoyuan]

Queryable passes the expression as-is to the provider and let the provider to decide about optimization. It shouldn't try to optimize any thing.

[LeaFrock]

It shouldn't try to optimize any thing.

Why shouldn't? Are there any cases of providers which decide to make Skip(0) a meaningful operation? Semantically speaking, it's just a no op. I prefer to think that it should be ignored by the very beginning.

Futhermore, let me take EF Core (for SQL Server) for example.

int n = ...;
var result = query.OrderBy(...).Skip(n).Take(10).ToList();

If n is 0, it still uses ROWINDEX among the sql instead of SELECT TOP(10) * from ... order by .... IMO, it's better to optimize Queryable than let downstream providers handle it.
However, it's just a discussion. Please let me know if anyone has any thoughts.

[TonyValenti]

I can see the advantage of optimizing out those two items. Ideally, a take(0) query should never even touch a database or collection.

I kind of think that a take(0) should return enumerable.empty or whatever the queryable equivalent is.

[acaly]

As @TonyValenti mentioned, this behavior is different from the System.Linq.Enumerable.Linq, which does not even touch the source if count is equal or less than zero (even the GetEnumerator() is not called)