It's all about the answers!

Ask a question

DNG: Overview about artifact with all outgoing and incoming links in all components

Ulrich Simon (91) | asked Jun 15, 9:35 a.m.

 I have Configuration Management enabled in DNG V7.0.2.

There are several Components, and a specific artifact of a module resides in different versions in these components. There are differente outgoing and incoming links to this artifact.
Is there a chance to get an overview about all outgoing and incoming links of this artifact in all components?

3 answers

permanent link
Bartosz Chrabski (3.1k11939) | answered Jun 15, 10:04 a.m.


OSLC query and interating through all the components should solve the problem.

There are two issues I see here:
  • It is custom development so it may be an issue
  • One component can have multiple wokring streams, so it may be component*streams iterations.

permanent link
Ian Barnard (780612) | answered Jun 15, 10:59 a.m.
edited Jun 15, 11:00 a.m.

The OSLC APIs certainly allow you to get at all the data, but that's complicated to code and then you have to visualise the information.

Another way is that the information about these links to different versions is in LQE if you use it, so using Engineering Insights (ENI) should be able to extract the information from the All Data source and visualise in a view using custom Artifact Elements and SPARQL, and then ENI does the visualisation work for you. I'm sure IBM Services team, or partners, could help you - contact your IBM Account Manager to discuss what you need. 

Bartosz Chrabski commented Jun 15, 3:35 p.m.


ENI may help but it will require coding in SPARQL to get this data too. OSLC wont generate additional license cost which is required by Engineering Insights.

Both solutions ENI and OSLC will solve Ulrich problem. Question is cost and skill they have.

Good Luck !

permanent link
Geoffrey Clemm (29.8k23035) | answered Jun 23, 1:12 a.m.

 It is sufficient to use the All Data LQE endpoint, and then use the Report Builder to build your query (no custom SPARQL should be required).   Note that you should not try to do any joins in the report (i.e.not query for more than one relationship) ... just query for the one relationship you are interested in, and select the appropriate sets of attributes of the source and target artifacts.

David Honey commented Jun 25, 9:54 a.m. | edited Jun 25, 9:55 a.m.

The problem with this is that you get the cartesian product across the versions of source and target artifact. For example, if there are 5 versions of a requirement each validated by a test case and there are 3 versions of that test case, you may end up with 15 rows in the report, rather than 3. This can expand exponentially as you add deeper traceability relationships.

Geoffrey Clemm commented Jun 25, 10:21 a.m.

Yes, this is why it is important not to query for more than one relationship in a single report (i.e. don't do "deeper traceability relations").   Also, whichever technique you use, you will want to post-process the results, to only keep rows with the distinct concept ID's.

Your answer

Register or to post your answer.