Linq查询非常慢

Question

I have a table that contains a list of builds. What I am trying to achieve is to group those builds by reference ID and take 5 or 10 latest (by build_date ) and see check their statuses. So I am currently doing the group by and returning them.

using (var ctx = new Entities())
{
    return ctx.Builds.Where(x => x.build_date > dateTime)
                .GroupBy(x => new
                {
                    x.configuration_id
                })
                .Select(s => new ConfigurationBuilds
                {
                    ConfId = s.Key.configuration_id,
                    Builds = s.Select(x => x).OrderByDescending(d => d.build_date).Take(count)
                })
                .ToList();
}

Later on in another method I will check all the statuses Builds items to see if they match any criteria, something like:

items.Where(build => build.Builds.Count() >= 9)
                    .Where(build => build.Builds.All(x => !x.build_status.Equals("SUCCESSFUL")))
                    .ToList();

The problem is with the first query where I do groupBy . That one is extremelly slow, how could I optimize that routine?

I createad a dummy routine to select all builds and group by statuses

var items = ctx.Builds.Where(x => x.build_date > dateTime).OrderByDescending(x=>x.build_id)
                    .GroupBy(x => new {
                        confID = x.configuration_id,
                        status = x.build_status
                        })
                .Select(s => new 
                    {
                        Id = s.Key.confID,
                        Status = s.Key.status,
                        Cnt = s.Count()
                    }).OrderBy(d=>d.Id)
                    .ToList();

This one on same data is blazing fast, but does not really do what I need. I need to: For each configuration group select n number of latest builds that match some criteria. How to do that?

Answer 1

A good way way is to determine what SQL it's generating is to use a tool like LINQpad . Once you get an idea of what the SQL looks like, you can use SQL tools like execution plan generation.

Or you can use SQL profiler to see what's going on with the SQL queries too.