The goal of solution is to archive data from HANA native database to Hadoop storage and access these archived data in HANA native. For this purpose DataTiering creates Virtual table and calculation view, it is only possible to create one calculation view for one DataProvider.
The created calculation view contains a union of a source table and a target table (Virtual table) in Hana native. The Hadoop is added to HANA native as a remote source through Smart Data Access (SDA). The Offloading will create a Virtual table, which is pointing to the Hadoop table. With this procedure we can access data from Hadoop in HANA native.
Features:
Support tablename longer then 16 characters
DP name HTAB*generated_hash* will be created. In main screen user can change view with
DP Description button to see name of Hana native table in description
Support table fields longer then 30 characters
The long field is cut or replaced with field with count (fieldname_count)
Relative condition
Custom target table name
After defining target storage, user can define custom target table name with mandatory prefix ZOFF* which is automatically added to the name
Possible to offload Hana native table with same name, but from different databases (schema)
if DP name is already defined, DP name will be automatically generated (HTAB*)
Set up before offloading:
You can check or set up storages in Storage management through the transaction /DVD/SM_SETUP.
Source storage - HANA Native Storage type SM_TRS_HNV
DB Connection Name - You can find an existing connection through the transaction DBCO. If there isn't any connection, please contact IT-Support.
Target storage - Hadoop cluster (set up by the IT-Support, except for the SDA parameters)
SDA parameters
If there aren't any SDA fields, open the content of the table /DVD/SM_STORTYP and check SDA_SUPPORT for the Hadoop storage type SM_TRS_MS.
Remote source name: Hadoop cluster
Database name: Database name
SDA Schema: Database schema of the primary DB
SDA Database connection: DB connection to external HANA database (transaction DBCO)
Smart Data Access (SDA): Remote source in source storage HANA database
Before you set up the target storage, you must check remote sources in HANA studio. According to the below figure you should fill out SDA parameters in the target storage (figure above).
If there isn't Hadoop as a remote source, please contact IT-Support, who can create remote sources in HANA native.
Only different steps for HANA native type are described here.
1. Create a DataProvider of the type HANATAB
The process of creating this type of DataProvider is the same as for other types. Enter the transaction /DVD/OFF and click Add DataProvider.
Mandatory fields are:
Source Storage - The field must be an Hana native storage (which you set up previously in /DVD/SM_SETUP)
DataProvider category
DataProvider name - Use search help to find a source table. Please note that the displayed tables are from external HANA
Target Storage - Hadoop (Hadoop storage with SDA parameters, which you set up previously in /DVD/SM_SETUP)
Target table name (optional) - table created in target storage, with mandatory prefix ZOFF*
Virtual table is created in this step and it is always saved into schema, where source table is defined.
2. Impact analysis
In this step DataTiering finds dependent objects of a DataProvider. After click on Execute Analysis the calculation view will be added. This means the calculation view will be generated in next step, therefore fields are empty.
3. Object adjustment
By the standard procedure, when the DataProvider is based on a InfoProvider, is in this step the object code adjusted in order to access archived data trough a lookup. However, by this process, when the DataProvider is based on a Hana table, we create by executing the adjustment the calculation view.
Generated Calculation view in Hana native:
