Often in SHR Content Development, we have come across the feedback that even to get data off an external Oracle or SQL Server database and create some reports in SHR, it requires quite a bit of xml authoring, Model development and Content development details.
Well, we’ve been listening and HP-Service Health Reporter (SHR), the promising cross domain reporting solution of HP now ushers in a simplified path for Content development [CDE Simplified…!] to build your data marts in a faster and simplified way….
Not necessitating explicit Data Model definitions or Workflow streams … (read on to find out how …)
The model driven content development in SHR has been through an Evolution…Let’s have a small recap on how Content development in SHR has evolved since its inception…
In SHR9.00/9.10…the “Model Automation” tool was used to develop content artifacts though it hadn’t been made generally available.
In SHR9.20… the new sophisticated content development platform (CDE) was used to develop the three layered content of SHR - with ETL, Domain and Application components – this was made available to Practitioners by and large to foster Content Development in SHR …
And now, in SHR 9.31… we present the “Simplified CDE” which offers the same three layer content pack creation, but in a smarter way for simple use cases.
You must now be wondering… Well, what is new in the Simplified CDE? The same three layered content seems to exist here as well …what’s new about it and more importantly, what’s simplified?
As you may all be aware, the CDE Platform of SHR9.20/9.30 enables the development of the ETL, Domain and Application content packs. However, for each of the content components, there is a list of input xml files that the content developer is expected to author.
For the domain component –a Model schema xml and ABC streams.
For the ETL- collection policies, transformation rules as required, stage rules and ABC streams.
For Application component- a logical model schema xml and report artifacts.
Each of the content source artifacts specific to each type of content pack, has to be compiled individually using CDE to create the installable package.
Therefore CDE provides a fully sophisticated process of defining the schema (either star or snowflake), hierarchy levels (for drill down or drill up), ETL processing, streams to control data flow etc.
But what’re the options in simpler use case scenarios where data from a source has to be loaded as is to data warehouse tables – even when end user is not interested in a sophisticated model definition, hierarchies etc.?!
Even in such cases, the content developer had been expected to take the longer route of authoring source artifacts thus far….
At this juncture, the Content Developer looks out for a smarter way of content development to address these simple use cases. And “CDE Simplified” dishes out the shortest and fastest way in achieving these!
Developing Content Using Simplified CDE:
Content development using “Simplified CDE” follows the below execution steps:
Create Content Folders
Define source definition xml as per the data source
Define Connection properties
Create Mapping File CSV
Create end to end content pack
Basically the Simplified CDE can be used to generate content pack for simple use cases sourcing data from Generic Databases, .csv files, topology sources like RtSM and from HP Operations Agents.
Depending on the source type, the content developer needs to author a single source definition file where the necessary queries and metadata would be defined.
The Simplified CDE would process these input files, and generate an intermediate file named “Mapping File. csv”.
The mapping file provides the details on how source tables and source columns have been mapped to the data warehouse tables and data warehouse columns. It also contains definition of required aggregations.
The Content developer can review and edit the default values present in the mapping file if need be.
The Simplified CDE takes the mapping file as input and generates the ETL, Domain and Application content pack components (with an auto-generated Business Objects Universe). Reports can be created off the generated universe and subsequently packaged as a .BIAR file.
We thus have the installable package of ETL, Domain and Application created in few minutes.