Hello, I need help with my data model: I have two tables, each containing a "tag" column. This column has similar values in both tables, and I would like to create a single filter that applies to the data from both tables, but it shows me two separate filters. How can I merge them?

Looking at your LDM screenshot, I can see you have two separate datasets — Live metrics and Queue metrics — each with their own

Tag

attribute. That's why you see two separate

Tag

filters in the dashboard filter picker. To have a single

Tag

filter that applies to both datasets, you need to extract

Tag

into a shared dimension dataset and reference it from both fact tables. Here's the approach: Solution: Create a Common Tag Dimension Dataset 1. Create a new

Tag

dimension dataset in your LDM (mapped to a dedicated tag table or view in your database). 2. Remove the

Tag

attribute from both

Live metrics

and

Queue metrics

datasets. 3. Create a relationship from the new

Tag

dimension dataset to each of the two fact datasets, using

Tag

as the primary key in the dimension and a foreign key in each fact table. [Create Relationships] Once both datasets reference the same

Tag

dimension, a single dashboard filter on

Tag

will apply to both. [Connections Between Datasets] > Important: For filters to work across datasets, the datasets must be logically connected in the LDM. As confirmed by GoodData support: "unless they are connected/joined in the data model, the filter does not apply." If the tables are in different data sources If

Live metrics

and

Queue metrics

come from different data sources, you can still use a shared dimension dataset — this is supported via the Multiple Data Sources in LDM feature. The shared

Tag

dimension must be present as a foreign key in each data source involved. [Multiple Data Sources] 👍 0 button 👎 0 button ℹ️ Tip: Mention @kapa.ai in the thread for follow-ups. 👀 Our Support team monitors all messages and will jump in if needed.

@kapa.ai How to create the primary key and foreign?

Based on your use case (creating a shared

Tag

dimension linked to both

Live metrics

and

Queue metrics

), here's how to set up primary keys and foreign keys in GoodData: Setting a Primary Key 1. In the LDM Modeler, select your

Tag

dimension dataset. 2. Click … (ellipsis) → Set primary key. 3. Select the

Tag

attribute and click Set key. [Set Primary Key] Creating a Relationship (Foreign Key) Once the primary key is set on the

Tag

dimension, link it to each fact dataset: 1. In the LDM Modeler, locate the

Tag

dimension (originating dataset) and one of your fact datasets, e.g.

Live metrics

(target dataset). 2. Click the blue dot on the right border of the

Tag

dataset and drag the arrow to the

Live metrics

dataset. 3. In the dialog that opens: ◦ The primary key (

Tag

) from the originating dataset is shown on the left. ◦ In the Reference Field dropdown, choose: ▪︎ Connect as new field — adds

Tag

as a new foreign key in

Live metrics

, or ▪︎ Existing dataset fields — reuse the existing

Tag

attribute already in

Live metrics

and make it the foreign key reference. 4. Click Connect. 5. Repeat for the

Queue metrics

dataset. [Create Relationships] > Tip: When reusing the existing

Tag

column already present in your fact tables, choose Existing dataset fields in step 3 — this avoids adding a duplicate field and keeps your model clean. After saving, both

Live metrics

and

Queue metrics

will reference the same

Tag

dimension, and a single dashboard filter on

Tag

will apply to both. [Primary Key Concepts] 👍 0 button 👎 0 button ℹ️ Tip: Mention @kapa.ai in the thread for follow-ups. 👀 Our Support team monitors all messages and will jump in if needed.

@kapa.ai do I need to create a view on my bigquery database ? Or only in the LDM?

Great question. Based on your situation, here's what you need to know: Do You Need a View in BigQuery? It depends on your data structure. Option A: Your

Tag

values already exist as a column in both tables If the

Tag

column is already present in both

Live metrics

and

Queue metrics

tables in BigQuery, you do not necessarily need to create a new view. You can: • Create a

Tag

dimension dataset in the LDM only. • Map it to one of the existing tables (or a dedicated tag table if you have one). • Use Existing dataset fields when creating the relationship to reuse the

Tag

column already in each fact table as the foreign key. Option B: Multiple datasets mapped to the same table If you need to map multiple LDM datasets to the same BigQuery table, be aware that publishing an LDM with multiple datasets mapped to the same table will fail. In that case, you must create separate views in BigQuery for each logical dataset. [Advanced Modelling Use Cases] BigQuery-Specific Note BigQuery does not support referential integrity (primary and foreign keys at the database level). This means GoodData cannot auto-detect relationships from the database schema. You need to define primary keys and foreign keys manually in the LDM. [BigQuery Additional Info] In summary: For your use case, you likely only need to make changes in the LDM (no new BigQuery views required), unless you end up needing to map multiple LDM datasets to the exact same BigQuery table. 👍 0 button 👎 0 button ℹ️ Tip: Mention @kapa.ai in the thread for follow-ups. 👀 Our Support team monitors all messages and will jump in if needed.