We want to leverage ML capabilities of Databricks for planning purpose. We have a simple input template in SAC and we want to understand if we need to replicate this plan data in any steps while SAC planning data is consumed in Databricks or this can be achieved via zero copy
Our current setup is shown below:
Question
If we want to use Databricks capabilities for planning data along with Seamless planning, is there a way to consume planning data in Databricks without replication? Or we have to replicate the data. I.e. Is above picture the expected architecture for such SAC Planning scenarios today or is there a supported zero-copy approach that allows Databricks to consume SAC Planning fact tables directly without creating ANM_SSA_DP (or another physical copy)?@VineetBakshi,@N1kh1l
Request clarification before answering.
Databricks Delta share ( Zero Copy) needs data to be in delta format ( parquet + delta logs). This format will be only available in Hana Data lake files and this is why only space with hana data lake is shown to you for delta share. So you would have to replicate the data into Hana Data lake file storage using replication flow. Also Zero copy is not a true zero copy and depends on what is the source. Its the source system responsibility to create data in delta format to support delta sharing which is nothing but temporary URLs to the files in the data lake itself. So you need to bring data to HDLFS in order to delta share.
Hope this helps !
Nikhil
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
| User | Count |
|---|---|
| 5 | |
| 4 | |
| 4 | |
| 3 | |
| 2 | |
| 2 | |
| 2 | |
| 2 | |
| 2 | |
| 2 |
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.