...
Exported file in a spreadsheet editor
Avoiding
...
duplicate records
It is a good practice to also include the primary key in the export definition, otherwise it can lead to duplicated duplicate records in the result of the export process. The duplicated duplicate records can occur because the dataset's columns and metric values needs to be joined and if there is no primary key, then the join is performed using all available columns. And if any of the columns contains some NULL values, the join process is not able to correctly match the records, because of the NULL values inequality.
...
In this case the metric values needs to be joined using columns branches.branch_name and branches.branch_city, because there is no primary key defined. And because the column city contains NULL value on the third row, the records are not correctly matched in the case of "Branch New York" and it results to duplicated duplicate records.
The result would look like this:
...