Forum Discussion
Replace Index with Values
Hello guys,
I'd need a hand with the model below. Table 1 shows uniques titles (movies, series, etc.) with its corresponding index "tconst". Table 2, shows uniques Names with its index "nconst". Table 3 shows every title (tconst), a person (nconst) and its role in the title.
Table 4 shows every title (tconst) and its director and writer (both are indexes nconst).
Which would be the best way to make a visual with: a title -> Name (from table 2) of Director -> Name (from table 2) of Writer?
Thanks!!
That's quite a handful.
This part needs cleaning up. Combine Location,Country and Region into one table, and clean up your "region" terminology. Your title region seems to be different from your geographic region.
Bidirectional 1:1 relationships most of the time mean you should fold the tables into a single table.
I think you have more than two fact tables. In fact (pun intended) you may not have fact tables at all. This gets close to a graph, and frankly is not something Power BI is good at. The Qlik associative model is more suited for such a scenario.
Can you try to formulate the main questions that your report should be answering for your intended audience? That may help to modify the data model in a meaningful way.
6 Replies
That's quite a handful.
This part needs cleaning up. Combine Location,Country and Region into one table, and clean up your "region" terminology. Your title region seems to be different from your geographic region.
Bidirectional 1:1 relationships most of the time mean you should fold the tables into a single table.
I think you have more than two fact tables. In fact (pun intended) you may not have fact tables at all. This gets close to a graph, and frankly is not something Power BI is good at. The Qlik associative model is more suited for such a scenario.
Can you try to formulate the main questions that your report should be answering for your intended audience? That may help to modify the data model in a meaningful way.
Many thanks for the help and advice lbendlin . I modify the questions in order to ease the help of the audience.
in a "normal" data model the filter direction is singular, from the 1 table ("dimension") to the * table ("fact").
Can you explain why you chose bidirectional?
Please post a sample pbix if possible.