WCF OData 性能改进

Question

I have following query which is taking a lot of time -

var allEmployees = (from e in context.Employees.Expand("Payroll/Customer")
.Expand("HR")
.Expand("Payroll1")                                                 
where e.IsActive
&& e.Payroll.EmployeeId== this.CurrentEmployee.EmployeeId
orderby e.Name
select e).ToArray();

Questions -

1. How can I improve the performance?
2. Which part of the query taking too much time, It can be Expand clause, where, orderby, select.
3. At last I am converting the result in Array. Can it impact the performance?
4. Do I have other alternatives, like Parallel.For, PLINQ etc. I am not sure about it.

Please suggest

Answer 1

The most important part is your Expand

No, It's not possible to LINQ over WCF data services with Parallel LINQ .

The best way is to index over foreign keys of relations between customers and human resources and payroll tables

ToArray wont takes too long

Use SQL profiler and SQL execution plan to improve your performance and find bottle necks

Consider creating method on your service layer and use Select method to return only columns you need

Another best practice is using JSON format instead of using XML Serializer. It will decrease your response size up to 70%.

Let me know if you want more information.

good luck

Answer 2

I know the question has been asked a few years ago, but if my advices can help someone, here they are:

1. How can I improve the performance? - Add Indexes for the columns you search or order by frequently. An index is ordering your column to than optimize query. You can do this in your DB https://docs.microsoft.com/en-us/sql/relational-databases/indexes/create-nonclustered-indexes or if you are using EF Core https://docs.microsoft.com/en-us/ef/core/modeling/indexes . So you can add indexes for Name and IsActive in Employees.
2. SQL Joins (.Expand) can slow down your performance, you should avoid them if possible. You can do so by flattening down the tables to avoid one-2-many/many-2-many relationships
3. I'd also recommend to debug this using Performance Profiler in Visual Studio (Debug -> Performance Profiler). You can launch it and query the DB and it will than show you the output documents with all queries. And than you can check this query via MSSQL Management Studio (if you are using MSSQL of course). For that you need to enable Live Query Statistics and see which instruction takes the longest.