Add Group By support for Aggregate Functions
We're really happy about the new aggregate functions that recently were announced. But without Group By support they fall short and will leave many disappointed. Please add Group By support as well.
Still in works, meanwhile, we recommend using the Spark connector: docs.microsoft.com/en-us/azure/cosmos-db/sp.. or docs.microsoft.com/en-us/azure/cosmos-db/ch.. to perform aggregates like Group By.
or Use Azure Stream Analytics (ASA).
Avijit Samanta commented
the existing aggregate functions just work on one field in a query. It does not work when use group by and multiple aggregate functions on a single query. the same works for cosmosdb mongodb api. the feature is really needed for flexible querying capabilities
GET. IT. TOGETHER.
Jason Chester commented
https://github.com/lmaccherone/documentdb-lumenize looks like it might solve some of the needs for aggregation while we wait for official support
Looking forward to this, in the meanwhile, there is a CosmosDB/Spark library I wrote to leverage this from .NET https://github.com/UnoSD/SparkSharp
Hello, Is there any update on this? By when it'll be released?
Niclas Rothman commented
You are recomending Azure Stream Analytics as a workaround, any example or can you outline
the solution you have in mind.
Jim Brown commented
This post is support for MongoDB inside of CosmosDB: https://azure.microsoft.com/en-in/blog/azure-cosmosdb-extends-support-for-mongodb-aggregation-pipeline-unique-indexes-and-more/
But there is a comment below which states that Group By is coming in the first half of 2018....finally!
Is there any update on this?
ram kumar commented
GroupBy is available in preview now as mentioned here:https://azure.microsoft.com/en-in/blog/azure-cosmosdb-extends-support-for-mongodb-aggregation-pipeline-unique-indexes-and-more/.
So when this is going to be live.
Ajit Sutreja commented
Is there an example/documentation of using Spark connector with C#/.NET SDK? According to this, https://docs.microsoft.com/en-us/azure/cosmos-db/spark-connector documentation you have to use JAVA or Python SDK.
Wow! it is almost 8 months since the request and there are 1327 votes for it, but we still don't have the future. I am really disappointed. :(
Hi, we are considering using MongoDB in CosmosDB, but reading the comments I guess MongoDB is not fully supported (at least, the aggregation framework). My question is, is there any place (web, forum..) where I can read if anything else is missing to analyse pro & const of using MongoDB vs DocumentDB in CosmosDB. Anyway, anyone here in this forum can shed some light on this? Thanks!
Nenad Marjanovic commented
We started using CosmosDB, we left for other solution, we tested again and we tried working with Azure support, then we moved to other platform again. A lot of vars from our application didn't work. As CosmosDB don't have all functions from native MongoDB in the end we have lost hope ... I don't know how we use this in production when we spend time being MS CosmosDB quality service. I can only say it's shame.
Rodd Harris commented
Was giving serious consideration to using this on a new project we're launching into -- however, had to drop it as serious once I realized there is no true aggregation (no group by). Can't be a serious solution without group by.
Andrew Choi commented
we definitely need this.
Is there a work around for this in the meantime? In SQL we could create a view?
Peter Molloy commented
Agree much better potential for data pipeline and analytical efficiency if more transformation libraries are provided at the language level. Top issue facing Cosmos today.
No GroupBy, No CosmosDB. Without a GroupBy feature the CosmosDB is not useful.
Critical fail for us as well :(
Critical fail. Aggregates without Group By is pretty much useless, leaving us to go through data RBAR. This needs to be fixed or else don't say you have aggregates.