Apache Pinot

```Hi team, I have JSON tags and metrics. the problem is that the data has METRIC JSON and the Pinot add schema does not allow me to specify metric field as json. How can I go ahead and define my schema ? :  
{ 
        "tags": { "real_name": "marvelmohinish99", "team": "TN4AF0V5W" },
        "metrics": { "salary": 8520, "performance": 22.5 },
        "mtime" : 1456249661342
}```

do you know set of metrics ahead of time?

Hi <@UDRJ7G85T>, Thanks for your reply. I am actually subscribing to  a kafka topic in which different kinds of metrics will come, both the tags and metrics are going to be heterogeneous.

`heterogeneous data in topic "USER_DATA"`


`{` 
        `"tags": { "real_name": "marvelmohinish99", "team": "TN4AF0V5W" },`
        `"metrics": { "salary": 8520, "performance": 22.5 },`
        `"mtime" : 1456249661342`
`}`


`{` 
        `"tags": { "ID": 2244, "team": "TN4AF0V5W", "SOFTWARE": "AND-1" },`
        `"metrics": { "usage": 33, "tables": 68 },`
        `"mtime" : 1456249661346`
`}`

<@UDRJ7G85T>: Your youtube video on linkedIn with apache pinot was awesome and inspiring!

for each "type" of message, I know the data that I need before hand.

ok, you have multiple options on how you want to model the schema in Pinot

Option 1
message_type| tags_json| metrics_json| mtime

i was thinking a of using option1, but was stuck at schema definition on Pinot interface. It allows me to define dimensions (tags in my case) as JSON but the interface does not allow me to define Metrics as a JSON field.

you can say metrics_json as dimension for now and it should work

sure, I will do that, Thanks a lot, Kishore

So I can perform aggregations on the metrics, even if I declare them as dimensions, right? just wanted to be sure.