Hi team, Two fold question here. 1. Am I able to...
# gd-beginners
Hi team, Two fold question here. 1. Am I able to have attributes that repeat across the LDM be filtered together? i.e a dashboard contains two visualisations pertaining to a Tourism Region in Dataset A, and the same Tourism Region in Dataset B. I would like to be able to filter on both Tourism Regions in both Datasets/Visualisations in the single dashboard using one single filter. For context, this problem extends out across many aggregation attributes including things like Suburb, State, Local Government Area and more, so it is not just the Tourism Region I am looking to filter across multiple datasets here. 2. In extension to this, are repeated attributes across different datasets of an LDM considered bad practice? The use case is that i have aggregate tables that represent different aggregations of data from much larger raw/source tables. I don't see a world where combining the aggregate tables into one aggregate table with all the aggregations together being possible, one due to inaccuracy of the aggregations and two due to computation speeds/costs of doing this for our business.
They have varying time aggregations as well.
Hi Paiwand. It’s not possible to filter different attributes using the same filter. If they have no relationship among each other, it doesn’t matter that they have the same name/data - by all intents and purposes, they are different attributes, and as such need to be filtered separately. Regarding question #2 - in theory, there’s nothing preventing you from doing that, but in practice it may cause some confusion when actually interacting with/using the data. You’re also held back a bit by point #1 above. I’m not sure how exactly your LDM is structured, but if you decide that this is the best way to represent your data, it shouldn’t be a problem.
Hi Paiwand, if the attributes have the same meaning, but are physically in different tables (which I believe is the case in your pre-aggregated datasets), you can actually make them work as one if you create a shared dataset from them. You just create new table/dataset with all the values and replace your original attributes with a reference to it. Then you can have a single filter for them and you can also combine data from different tables across this attribute. I believe it is exactly as point #3 described here: https://university.gooddata.com/tutorials/data-modeling/logical-data-model-basic-rules-of-data-modelling/ Just please note that to make everything work as expected, the shared table need to have all values that could exist in any of the tables referencing it. Including for example an empty value (in case your attribute is not mandatory).
🙌 1