As of March 30th, 2018, users of Watson Analytics Plus and Professional will notice a small change in the product. Specifically, users who leverage Data Connections from a Relational Database will notice that the process by which their data is imported has been changed.
The new process to import data from a relational data source is a necessary change to the product, and after a short adjustment period, will be seen as an improvement to Watson Analytics as we know it. Users and Administrators alike should be aware of the specifics of the upcoming change, as some work will be required of them up front in order to ensure ongoing functionality of existing data connections, and to avoid confusion when setting up new connections in the future.
The new Data Connection process utilizes an IBM tool called Data Refinery. Unlike the previously used tool, Data Refinery will not be embedded in Watson Analytics. This means that administrators will need to register their IBM id with IBM Data Refinery outside of Watson Analytics, then when ready, will use this tool to initiate the ‘Data Flow’ from the relational database in question to Watson Analytics via IBM Data Refinery.
Administrators must take the following preliminary steps to allow IBM Data Refinery to function with their instance of Watson Analytics:
- Access IBM Data Refinery
- Sign in using your existing IBM ID
- Register for IBM Watson Knowledge Catalog
- Select Watson Analytics your ‘target’ for the Data Refinery output
Once configured, administrators are free to use IBM Data Refinery to import data from a relational database by taking the following steps:
- Launch Data Refinery
- Create the extract process or ‘Data Flow‘ that selects and cleans the data as needed
- Load the data into Watson Analytics
It is important to note that IBM Data Refinery will recognize existing asset names in your instance of Watson Analytics. So, if you are updating an existing Watson Analytics data set, simply use the same data set name in Data Refinery — this will update the data set with the new data outlined in the Data Flow. If you’re ingesting brand new data into Watson Analytics, just provide a name, and execute the process in Data Refinery.
We understanding this process requires some additional work on your part to make your data flow properly. We are confident that after the initial work is complete, users will find that leveraging IBM Data Refinery to import data will simplify the process while also allowing users to interactively discover, cleanse and transform their data to a greater extend than ever before.
Thank you for your continued business.
For a more detailed discussion of how to configure IBM Data Refinery with Watson Analytics, see our recent step by step document which walks through the process in depth.