Hello GD team, in a GD workspace, let's suppose a LDM can have multiple clusters (each cluster = a group of tables linked together). Suppose we have a table to be used in multiple clusters, we're wondering if a table can be imported multiple times? (it seems like we could only use a table once in the LDM) 😅 May I also get GD advice on 1) how to resolve this, and 2) other clients with such a need and how did they go about resolving this? cc @Jennifer Chue

Hi Alson, this should be possible to achieve in GoodData LDM if you need that, but maybe you can’t do it by simply drag and dropping in the modeller. But first - if the table is common for multiple of your table clusters would it makes sense to view and filter the data by it across the clusters (i.e. if the table was “product” have “sales by product” and “support tickets by product” analyses), If yes, the you can also consider keeping it in the model just once and connecting it to your clusters with references (arrows). Just make sure to stick to the modelling best practices (esp. #5) while doing so. But if you really prefer to have it separate also in GoodData for some reason, you can create separate datasets and map them to the same physical table. Dragging and dropping the table would not work (it will probably try to stick to the existing dataset), • You can create new dataset manually (drag empty dataset from the menu) and add the columns and map it to the same physical table in the mapping dialog. • Or if that would be a lot of columns to do manually ◦ export your current model to a JSON file (under the tree dots menu in top right corner) and copy/paste the existing dataset and its mapping as another dataset in the file (just make sure to change all the identifiers to keep them unique). ◦ or you can possibly create a 1:1 view in your database and then use it with the drag&drop - that will work One more tip for naming the fields - it might be a good idea to also have different names for the same columns in different datasets. Unlike identifiers, this is not mandatory, but it will make your life easier if you use them in metrics etc. - you will always know which one it is.

Thanks a lot Michal, greatly appreciate your detailed answer. I'll look further into the link that you've provided. And I've actually tried the method which you mentioned about using an empty dataset and map it to that physical table which has worked well for me. True enough it was quite a hassle adding those columns manually again, thanks for suggesting the other alternatives. This has helped to answer our qn and change our approach on doing things 😄