The Product Availability Matrix of SDI does not show any adapters, so seems not to be the case.
Writing new adapters yourself is not big deal, I have written one for Azure Data Lake to read Json and Avro files. Given that Cosmos DB does not have a schema being a NoSQL DB, the question of that capabilities would be important. Do you want to see the document just as one large binary object? Do you want to apply where clauses? If fields are used, what structure should the virtual table have when there is no schema in the source?
For the Data Lake adapter I had the same questions to answer as it is a large files system with Json documents in it. What we did there is to define a view on top of the Json and one view is one virtual table. Now we do have a structure for the virtual table and thus it can be queried in Hana via SQL.
Further more the question if realtime changes ("subscription") should be supported as well. For the Azure Data Lake we used Event Grid to support that. For Cosmos DB I am not sure it is even possible. Do you have an idea?
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.