Hitting the 2100 parameter limit (SQL Server) when using Contains()

Viewed 78188
from f in CUSTOMERS
where depts.Contains(f.DEPT_ID)
select f.NAME

depts is a list (IEnumerable<int>) of department ids

This query works fine until you pass a large list (say around 3000 dept ids) .. then I get this error:

The incoming tabular data stream (TDS) remote procedure call (RPC) protocol stream is incorrect. Too many parameters were provided in this RPC request. The maximum is 2100.

I changed my query to:

var dept_ids = string.Join(" ", depts.ToStringArray());
from f in CUSTOMERS
where dept_ids.IndexOf(Convert.ToString(f.DEPT_id)) != -1
select f.NAME

using IndexOf() fixed the error but made the query slow. Is there any other way to solve this? thanks so much.

5 Answers

Why not write the query in sql and attach your entity?

It's been awhile since I worked in Linq, but here goes:

IQuery q = Session.CreateQuery(@"
         select * 
         from customerTable f
         where f.DEPT_id in (" + string.Join(",", depts.ToStringArray()) + ")");
q.AttachEntity(CUSTOMER);

Of course, you will need to protect against injection, but that shouldn't be too hard.

You will want to check out the LINQKit project since within there somewhere is a technique for batching up such statements to solve this issue. I believe the idea is to use the PredicateBuilder to break the local collection into smaller chuncks but I haven't reviewed the solution in detail because I've instead been looking for a more natural way to handle this.

Unfortunately it appears from Microsoft's response to my suggestion to fix this behavior that there are no plans set to have this addressed for .NET Framework 4.0 or even subsequent service packs.

https://connect.microsoft.com/VisualStudio/feedback/ViewFeedback.aspx?FeedbackID=475984

UPDATE:

I've opened up some discussion regarding whether this was going to be fixed for LINQ to SQL or the ADO.NET Entity Framework on the MSDN forums. Please see these posts for more information regarding these topics and to see the temporary workaround that I've come up with using XML and a SQL UDF.

Related