Redis OM dotnet - Group by + Count

Question

I have started using Redis Stack recently with dotnet. I am however having issues to figure how to achieve the following aggregation.

I have a list of sport events, each of them, among other properties, has SportCode and Status. I am trying to group my entities in a way I can get the following result:

[
  "sport": "SOCCER",
  "eventsCount": 400,
  "liveEventsCount": 12
]

So far I've tried 2 approaches:

1. Use RedisAggregationSet<TEntity> - this gives me the groups, but I did not see a way to get count of total items per group or count of items that match a condition (to get the live events)
2. Use IRedisCollection<TEntity> - grouping here always throws a JSON serialization error, even after I've changed all my records in Redis to be saves as hash, instead of json.

Answer 1

Assuming you have a model like:

[Document(StorageType = StorageType.Json)]
public class SportEvent
{
    [Indexed(Aggregatable = true)]
    public string SportName { get; set; }
    
    [Indexed(Aggregatable = true)]
    public string EventType { get; set; }
}

This:

var res = aggregations.GroupBy(x => new { x.RecordShell.EventType, x.RecordShell.SportName }).CountGroupMembers();
foreach (var result in res)
{
    Console.WriteLine($"Sport Name: {result["SportName"]}, Event Type: {result["EventType"]}, Group Size: {result["COUNT"]}" );
}

Will create a result set that has the information that you need for the intersecting groups (where you are grouping by both attributes). However if you need to group things individually e.g. first by sport and then by event type and sport, you could either run two seperate aggregations (one for each groupBy clause) or you could run a single GroupBy and then sort it out in your app.