Creating a Calculation View from the Union of In-Memory and Extended Tables

We will be creating a calculation view called CA_UNION, which will be a union between ORDERS_CS and ORDERS_DT. If these tables have not been set up, follow the instructions in the Prerequisites.

To start off, right click tutorial_All_Calc package and click New > Calculation View

In the subsequent screen, fill out the values for your calculation view. Name it CA_UNION, and make sure it’s type is Graphical, and has a data category of Cube. Do not check the “With Star Join” box. Click Finish to continue.

You should now see the graphical editor for your new calculation view. To start building our calculation view, drag one union node and two projection nodes onto the screen. Rename one of the projection nodes to “Hot_data” and the other to “Warm_data”.

Note: To rename a node, click on their title.

After renaming, your diagram should look like this:

We will now add the ORDERS_CS table as an element in Hot_data and the ORDERS_DT table as an element in Warm_data. To do this, navigate to where your ORDERS_CS and ORDERS_DT tables are in the left Systems panel, and drag/drop them into Hot_data and Warm_data, respectively.

Note: If you have followed the tutorials, the tables should be in <SID> (TPCH) > Catalog > TPCH > Tables.

Select the Warm_data node, and you will see the details show up under the “Details” panel on the right hand side. Right click the top of the diagram and click “Add All to Output”.

After doing this, all the gray circles will become orange. This means those columns will be sent to the next node.

Repeat this process with the Hot_data node.

Now, connect the Hot_data node and the Warm_data node to Union_1. To do this, press your cursor over the gray dot on the top of one node, and drag it over to the gray dot of another node.

Select Union_1 to see its details under the “Details” panel. You will see 2 sections - one called Source, and another called Target. We will need to create source and target mappings for the tables we are creating a union for. Under source, expand Warm_data. Select all columns (hold down ctrl), left click, and choose Add to Target.

This will add these columns to the target section, along with mapping lines. Next, we will be mapping the columns in Hot_data as well. This can either be done manually, or be using the “Auto Map By Name” button, since both tables share the same columns. We will be doing the latter way.

Back in the Scenario panel, join Union_1 and Aggregation together by clicking on the gray dot at the top of the Union_1 node and dragging it over to the gray dot at the bottom of the Aggregation node.

Now, select Aggregation to edit it in the Details panel.

In the Details panel, you will see the columns which have been passed from Union_1. For every column except O_TOTAL_PRICE, right click and select “Add to Output”.

These columns will now be outputted as attribute types. For O_TOTALPRICE, right click and select “Add as Aggregated Column” instead. This will set it as an output column of type measure.

This is important, because all calculation views of type Cube require at least one measure.

Finally, select Semantics from the Scenario panel, and view the details in the Details panel. Ensure that the Aggregation for O_TOTALPRICE is set to Max, and toggle the Type so that all columns other than O_TOTALPRICE are of type attribute of a blue diamond.

The aggregation defines how O_TOTALPRICE will be joined. Since there shouldn’t be any duplicate rows between ORDERS_CS and ORDERS_DT, it does not really matter.

We have now completed our calculation view, which created a union between both in-memory and extended tables.

Navigate to tutorial_All_Calc in Systems. Under tutorial_All_Calc > Calculation Views, right click CA_UNION and click Activate.