Dimensions and primary keys

In the Google Analytics data source, the set of dimensions you select affects how Panoply generates the primary key for records retrieved.

This is because, for each dimension you add or remove, you're basically affecting how the source's data is aggregated and ingested into Panoply. This is especially true when adding more dimensions. The more you add, the more detailed or varied the data to collect. This, in turn, results in the creation of a distinct record with a primary key of its own.

As an example, we'll use data taken from a website that analyzes its number of users per country. We limit the case to two countries only, namely, Brazil and Spain.

First, here's the result after running the data collection using only the ga:country dimension:

Last, we make a change by breaking it down per city. Here's the result after recollecting the data using both ga:country and ga:city dimensions:

The data (number of users) from the two results remain the same. It's just that they are aggregated differently because of the dimensions specified. But one thing is constant, for each distinct record brought about by the change in dimensions, a unique identifier is provided for it.

To conclude, just keep in mind that any change you make to the list of dimensions will affect the value of the primary key. If you plan to change them, it's recommended to either ingest the data to a new table or delete the records from the existing table first before recollecting the data.