Child pages
  • SAP HANA-optimized BW Objects - Infocubes
Skip to end of metadata
Go to start of metadata

Important properties

 

Flat cube

A HANA optimized infocube is a flat cube. This means that only the package dimension is still created as a dimension table. The SIDs of all other characteristics are stored directly in the fact table. An overview about the new enhanced star schema can be found here in the documentation. In addition, also the E fact table is gone.

Whether an infocube is already optimized can be verified using transaction RSA1:

 

In addition, the path 'Extras' -> 'Information (logs/status)' -> 'Dictionary / DB Status' would only show 2 tables, the F table and the package dimension:

 

Partitioning

The F tables of HANA optimized infocubes are partitioned. Details regarding how partitioning works on HANA side can be found in the documentation here. Information about the current partition layout of a table can be retrieved using the following SQL statements:

get partition specification
select table_name, partition_spec from tables where table_name = '/BIC/FSTPEOLAP3'

 

The following SQL statement can be used to retrieve detailed information regarding where partitions are created:

 

select mcp.part_id, mtl.host, mlhc.indexserver_actual_role 
from m_cs_partitions mcp 
	join      
	m_table_locations mtl on mcp.part_id = mtl.part_id and                    
				 mcp.table_name = mtl.table_name               
	join          
	m_landscape_host_configuration mlhc on mlhc.host = mtl.host 
where mcp.table_name = '/BIC/FSTPEOLAP3'; 

 

As expected on a BW system, the data is distributed only over the slave nodes.

In addition, HANA Studio offers a way to directly access the current distribution of partitions via this path: Expand "Catalog" Folder -> Right click on table -> Open Definition -> Runtime Information

Partitioning - Checks

In order to verify whether the current partition specification confirms to the expected value for the respective BW object, the report RSDU_TABLE_CONSISTENCY can be used. Before using this report, it is important to always import the latest corrections relevant to this report as outlined in SAP note 1937062. The relevant check option in this case is CL_SCEN_PARTITION_SPEC:

In addition to the partition specification, also the location of the partitions can be checked. This check is part of the expert options that are described in SAP note 1937062. This might be useful in cases where the partition specification itself is correct, but partitions are located on the master node for whatever reason:

However, it is important to note that any problems regarding distribution of partitions should only be rectified using the HANA landscape redistribution feature which will ensure an optimal distribution of data. See SAP note 1908075 for additional information.
  • No labels