SAP launched its new data warehouse product which is completely in cloud called SAP Data Warehouse Cloud, with growing popularity of Snowflake and AWS RedShift, it was inevitable for SAP to have a cloud warehouse product. SAP has been making investment in BW space for long, first with BW on HANA and then BW/4HANA, the new offering takes a new route with pure cloud play.
With huge upfront cost for on-Premise solutions, Warehouse on Cloud seems like a right strategy to have. Pay as you go has become norm with many industry solutions and SAP is headed in the same direction with SAP Cloud Platform and SAP Data Warehouse in Cloud.
The new system should be available for free trial , this blog is to give you a sneak peek of what lies under the hood. It can be a guiding document when you get access to system ( However SAP themselves have excellent video guide for the same ) .
This is the path we will be following with this blog, We will create a Space, Build a Data View and Visualize the Data. Lets get started!
This is the Home screen you see post login. I really liked the clean UI. We see various option on left-hand side and in the middle we have buttons with which we can perform various actions.
Once we click on the top button on left-hand, we see extended panel with the text. Here we have various option available. (Some are still work in progress).
Lets start with first logical step. Space, SAP Data Warehouse Cloud has concept of ‘Space’ which is nothing but logical segregation of data, For example we can have Space named as Marketing, Sales etc. Basically you build container and put data/model into it and let respective business team explore it.
When you click on Space Management button you will get to see current Spaces and option to create a new one. We can also see the data usage.
To create Space , Click on + sign and give a unique name.
Under the new Space, we can assign memory and members (who will have access to that space), Here we have configured 1 GB space (by default) and assigned it to my user id. You can do a connection assignment from various sources but to keep it simple we have skipped that.
Next we will click on the Data builder option to create our data model/view. Click on Data Builder and choose the Space you want to build data model/view in.
We see multiple options here , We can create new ER model, New Table, New Graphical/SQL View, or Import/Upload a file.
We will be loading the data via CSV file (Max Limit is 20MB per file). We have Sales and Sales Item files, we will be loading both the files one by one by clicking on Import/Upload file.
You will see the preview of data and you can change the data type as you deem fit. Strangely , system auto assign all the columns a same data type i.e. String (5000), We changed the data type of some of the measures to decimal data type.
We also loaded Sales Item file. Next step is to build model/view on top of the data. To build a model/view you can click on New Graphical View.
On Initial screen of Graphical View, On left hand side you will see the tables we have loaded under Repository tab, You will see all your external source meta-data under Sources tab (we don’t have any connection created for time being).
To start the modelling, You can simply drag the table to the canvas. Here we have dragged SalesOrders table on the canvas.
Once you drop the table, the output table automatically gets created and we also see the number of columns in both the tables (17). We also see options like Filter, Rename/Hide Column, Formula, Join Suggestion and Table Preview, all are self explanatory.
Next step is to create a join between Sales and Sale item data, for that we have dropped the item table on top of SalesOrders table.
System automatically assign a join between these 2 tables.
To see the Join details, click on the join button and then the Details button on right hand corner. We can see various join types are available and we can choose as per our need. We can also see that system has assigned join link between SALSESORDERID fields, we can change it, if required.
Box next to join is projection. In Projection, System automatically hides the duplicate fields, we can make them visible if we want but system will throw an error if you have duplicate columns (if you want to keep both the columns, you can just rename one of the columns).
Good thing about this model view is, It is aligned left-to-right and not bottom-to-top ( hence more easy on eyes in my opinion) and here you can see data preview at each step.
We will be giving a business name to output table to make it more user friendly and convert the type to Fact because we want to consume it in our reporting layer. Click on output box and then details button.
We will be converting some of the attributes like Gross Amount, Tax Amount, Net Amount to Measure.
To make it more contextual we can add the text in Business Purpose section, as shown below.
We are ready with our Graphical View, we will save it.
And to make it available to reporting layer, we will deploy it as well.
We will go ahead and create a story on top of it, we have SAC in-built in SDWC .
Click on Create Story and then add the data model/view.
Post data binding we will create a Chart.
We will simply select Gross amount in measure .
And select Sales Org as dimension and our chart is ready. Similarly you can visualize more data in SAC.
Thank you for reading till here, we just loaded the data in SAP’s latest warehouse solution on cloud and built a small visualization on top of it in SAC.
I really liked the intuitiveness of product and minimalistic UI has enhanced the UX quite a lot. Functionally, product is still in primitive stage, It looks like enhanced version of BODS, with addition of db in the back end and SAC on top of it. Many warehouse features are still missing, Like OLAP options, Scheduling of the data, Master data management etc. but lot of new things are still in pipeline and hopefully it will be ready by the time it is available for trial.
Right now SAP is positioning it as a service on top of your existing warehouse solution and not as an alternate warehouse option but once SDWC has all the warehouse feature, it will be difficult to convince customers to use both the warehouse tools together in the future.