While working on some updates to a table I was able to set up linked columns within the same table. I’m working on this to identify if we can establish parent/child relationships between different assemblies, so that e.g. I can say which IDs of Part A were used in Part B and retain that relationship for future reference.
I just wanted to confirm that this was intended functionality, since the vast majority of times I’ve seen Linked Columns referenced, it has been between two different tables, and not within the same table.
Hey @David2 ! The linked record feature in Tulip is one of the most hotly debated features that we have (both within Tulip and amongst the user-base). Despite its shortcomings, I still believe it offers a lot of valuable utility. It remains the only method to store actual arrays (not delimited strings) within table records.
The method you are describing is one of the things I would argue that linked records is well suited for, and there is no better alternative. Devout linked-record anti-enthusiasts might argue that you could record the lineage with IDs and forgo the linked record entirely. While this may be true, I think linked records make it simpler to search the lineage because you can simply select an assembly record and filter the same table where ID “Is in” that same assembly record’s children array. If you were to not use linked records, you would need to query that children array.
Long story short, I personally support this method and approach as a good use of linked records. I also support the use of linked records in many-to-one relationships where I want to display reference information (such as item images) in a secondary interactive table which references the item (such as a work order table).
Thanks for the response! One note for us, and why we’ll probably go with the linked record over an array, is that the inclusion of a relationship between assemblies is important information, and we’d need to be able to verify if the relationship was changed from either end of the linkage.
There’s definitely some room for conversation on the pros and cons of establishing relationships between/within tables using pure relational algebra, or using linked records. I think a lot of this is driven by the fact that the robustness of querying is limited since it can’t account for record linkages. Linked records would be much more powerful if it was readily possible to query records from Table A based on the status of associated records in Table B. Right now this requires some cleverness with aggregations and holder variables.
