I have a knowledge graph consisting of data virtualized from Databricks (i.e., a virtual graph in
.sms2 format) and an associated named graph containing properties and inference rules (in
Can I combine these two files into a
.stardogdesigner file and load into Stardog Designer?
((I realize that this is the opposite of "no-code" environment. However, I am going back and forth between Designer and Studio because the designer doesn't 100% produce what I want (but it is a great way to learn the ropes from zero).)
The short answer is no. We currently do not support importing existing mapping files (
.sms2) or existing mappings in Stardog to Stardog Designer. These are capabilities we plan to add in the future.
.stardogdesigner file is machine-readable and cannot be built except as an export from Designer.
I appreciate the feedback and am glad you are finding the tools provided useful, including identifying some of their limitations! Out of curiosity, what mapping capabilities are you looking for that are not supported in Designer?
An example of mapping capabilities limitation: when loading data from virtual sources, say several tables have the same attribute name (i.e., column name). If these tables are added together the attribute names are identified as the same and grouped together. However, if I add these table one at a time each the attribute names are deemed as different and
_2, and so on are added to the attribute. This means iterative building has a hurdle built into the process.
Additionally, the rule creation UI is limited. If I wanted to create a more complex IF statement (e.g., a sub statement with an aggregate and a group by clause), I would need to code that specifically. I can code that but each time I update the information in the Designer file I have to copy and paste that information. I would say this leads me to see Designer as testing grounds for development while development would take place in Studio.
If there are areas I could provide clarity, please let me know.