Introduction and Scenario

Part 1 of this series on Action Links had us creating a simple navigate action and left us in suspense with the promise of a new means by which to filter one report from another. Now time to deliver! The scenario: say we’ve got a company dashboard that is pretty locked down in terms of adding any new user-facing content, and as a developer you occupy more of an analyst role where you can build analyses all day, but dashboards are a no go. So how then can we drive some of our own detail analyses from dashboard selections? The answer? Hidden driving documents! In a nutshell, these documents store selected dashboad values and act as a filter for our detail analyses, utilizing OBIEE’s “filter on the results of another analysis” filter option. It should be noted that we could simply set the target analysis’s filters to be equal to the presentation variables generated by the source dashboard prompt. However, the following technique is simply to illustrate a possible technique, say when you’re going between two different subject areas, that might be applicable in your current situation or at least spark an idea to get you headed in the right direction towards a possible solution. So, let’s start by getting the pieces to our puzzle together.

Start with the dashboard and prompt. We are going to navigate from a monthly sales trend to a detail report displaying product sales detail.

The Solution

Now that we’ve determined the dashboard analysis from which we want to navigate, we can set up our driving document. Again, the overall concept of a driving document is to act as an intermediary container for selected values that then get passed from one report to another. So let’s set this up. Given that our trend is prompted on Year, Company, Region, we need to place three dummy columns on our driving report to store the presentation variables generated by each prompt. I used the same three columns in my driving report for consistency, however any column will do. Note that the columns in the picture have custom names already are not the native columns from their respective tables.

Edit each column to store the presentation variables generated by your dashboard prompt. If the dashboard prompt does not have a presentation variable for each column prompt, go ahead and get those set up before completing this step. My year column variable in this example is appropriately titled SALES_YEAR. Be sure to put single quotes around your variable as in ‘@{SALES_YEAR}’ when entering it in the Edit Column dialogue.

You can now save your driving report and place it on your dashboard. Right away you should see that your driving report picks up the values in each column prompt. Note the use of a default value {2012}. Using these is ideal so that if we want to edit our driving document, we at least have some values to pass it, as opposed to it throwing an error on the Results tab.

Now for some bad news. You might have noticed an obvious limitation to this approach, which is the driving report’s limited selection of only one column value from each column prompt. Nevertheless, let’s run this thing through to the end and see what happens. The next piece of the puzzle is setting up your target analysis.

Our analyst in this scenario wants to set up two reports to be driven off of our dashboard selection. The first will be a top 10 report for products based on the selected month. The second will be another top 10, but for sales reps instead. Navigating to our desired subject area, let’s bring over the needed columns, formatting as needed. In this example, we’re going to include a custom calc which gives us the percent variance to the prior period which has also been conditionally formatted to indicate positive or negative growth.
We have placed the columns necessary to build each report and also added filters, as seen in the proceeding picture.

Now comes the crux of this neat trick. We need to add the proper filters to the analysis so that any values we pick from both the dashboard prompt and the source analysis itself will drive our target analysis.

In the target report, we added a filter for each column that we’d like our analysis filtered on and likewise, comes from our source report.

To set this filter, simply apply the ‘is based on results of another analysis’ filter to each column. Then, browse for the driving report we made and apply the ‘is equal to any’ Relationship option. As our driving report values will only have one value for each of the filtered columns, our target report columns should filter on each of these. Lastly, use the drop down to choose the corresponding column from the source analysis. In the following pic we are selecting the  Region column from our driving report so that our target report will be filtered on whatever Region is selected on our source dashboard.

As the last steps in our target analysis, make sure you’ve set all the filter columns to respond to the driver analysis and that you select the ‘is prompted’ filter for any additional dimension columns on your report that are not driven by a dashboard prompt. In the example, this is going to be the Month column. This will ensure that our target analysis “listens” to whatever month is selected in our source analysis, that is, our sales trend.

Don’t worry if your Results tab shows no values as this is expected behavior (we haven’t sent any values over yet!). If you’d like to set it up so that your target report actually yields results without clicking on the action link, simply apply default values to the presentation variables in your driving document, as described above in the Year example. The last piece of this puzzle is finally assigning the action link to our respective column. Follow the instructions outlined in the first part of this post in order to navigate to our target analysis. In our example, this is going to be our Sales column.

Now that we’ve got all the pieces of the puzzle, let’s put it all together and see what happens. Clicking on any sales point on our trend line, we should be taken to the target analysis which should be filtered on both our dashboard prompt values as well as our Month. In this example, I’ve applied a filters object to the target analysis to make sure the target analysis responded to all applied filters.

Looks like everything works great! At this point you can hide the driver report on your dashboard and get rid of the filters object on your target analysis. I’d like to hear about some neat ways people have used this technique or modified it to suit their particular business/use case. Feel free to post in the comments! Stay tuned for part three of this series on Action Links where we go in depth about the GoURL syntax and how it can be a powerful and customizable tool in navigating to web pages and OBIEE 12c content.

The post Action Links in OBIEE 12c – Part 2 appeared first on Rittman Mead Consulting.