Aggregation query: `toList` reducer doesn't generate the expected list

[andy-stark-redis]

NRedisStack Version: current master branch

Redis Stack Version: 7.2.4

Description:
I attempted to write a code sample for the aggregation query doc page. Under Grouping without aggregation, there is an example that uses the toList reducer. The equivalent code for jedis and redis-py works OK but the NRedisStack code shown below doesn't actually add any output from the reducer:

AggregationResult res4 = db.FT().Aggregate(
            "idx:bicycle",
            new AggregationRequest("*")
                .Load(new FieldName("__key"))
                .GroupBy(
                    "@condition",
                    Reducers.ToList("__key").As("bicycles")
                )
        );
        Console.WriteLine(res4.TotalResults);   // >>> 3

        for (int i = 0; i < res4.TotalResults; i++)
        {
            Row res4Row = res4.GetRow(i);
            var res = $"Condition: {res4Row["condition"]}, Bicycles: {res4Row["bicycles"]}";
            Console.WriteLine(res);
        }
        // Should be an array of bicycle:x at the end of each line but nothing appears after "Bicycles:"

        // >>> Condition: refurbished, Bicycles: 
        // >>> Condition: used, Bicycles: 
        // >>> Condition: new, Bicycles: 

The "bicycles" index and the JSON objects are working correctly, btw, and are used elsewhere successfully in the same page. There appears to be an old issue in the StackExchange repo that is similar but I don't know for sure if it is referring to the same problem.

[schutztj]

The issue here is that the AggregationResult improperly parses the response:

internal AggregationResult(RedisResult result, long cursorId = -1L)
    {
        RedisResult[] array = (RedisResult[]?)result;
        _results = new Dictionary<string, RedisValue>[array.Length - 1];
        for (int i = 1; i < array.Length; i++)
        {
            RedisResult[] array2 = (RedisResult[]?)array[i];
            Dictionary<string, RedisValue> dictionary = new Dictionary<string, RedisValue>();
            int num = 0;
            while (num < array2.Length)
            {
                string key = (string?)array2[num++];
                RedisResult redisResult = array2[num++];
                if (redisResult.Type != ResultType.Array)
                {
                    dictionary.Add(key, (RedisValue)redisResult);
                }
            }

            _results[i - 1] = dictionary;
        }

        TotalResults = _results.Length;
        CursorId = cursorId;
    }

The RedisResult type for the TOLIST reducer is a BulkString array meaning we drop it from the results dictionary because of this check:

if (redisResult.Type != ResultType.Array)
{
   dictionary.Add(key, (RedisValue)redisResult);
 }

I've had to manually add parsing code combined with ExecuteAsync in order to handle these aggregations correctly.

[atakavci]

hi @schutztj , let me try to fix this and see if it resolves your case as well. we have an open pr for this now.