I performed my work at SAP Hungary Ltd. in the PLM Recipe Management team. In my thesis I present my own data models in SAP HANA database and describe the establishment of the relationship between HANA and analytical tools.
I prepared a possible modeling of the main components of the PLM recipe object, the header data and the formula. During development I tried to implement realistic models. I took available all the relevant information, the link between the tables were formed on this basis, in order to ensure the re-use the ID fields were connected to the output.
Significant part of my work is to achieve the recipe variant management. A number of version control methods are known in the system. I have used the opportunity given by the Change Management and the primary output specification of recipe modifications to follow the three models implemented accordingly. The first model is based on the simple idea, that the interval defined by the two input parameters iteratively call the basic model. The idea is not effective, so in the second model I did not use cycle, I filtered the result set, instead of costly inner join, external relations are defined, so I was able to implement a more efficient model. The third model is able to threat the inherited objects. In this case, due to a defect in the environment, I was able to solve the problem by using a cursor, which also slows down the performance. Significant enhancements caused by execution on the SQL Engine. Due to this, indirectly the run time of second model could be reduced to 1 second.
Building end-to-end communication between the HANA database and client applications is also a part of my work. HANA is typically optimized for data access via ODBC or JDBC protocols and SQL queries, but also supported by MDX language and ODBO interface multi-dimensional data service. After creating a successful relationship, objects in the HANA database will be available for clients, using devices further analyzes, reports can be created.