Classifications work like lookup tables that are evaluated whenever a report runs. It does not matter when a Classification value has been created or when the data has been tracked, as long as there is a lookup when the report is executed it will respect the Classification.
We have two ways to populate the lookup tables: First, we can use the Classification Importer through the frontend, FTP upload, or API. Second, we can use the Classification Rule Builder. But what does that actually do?
Classification Rule Builder does a few things:
- It queries for tracked values for the date range we set in the interface. If we select 6 months, it will basically run a report for the last 6 months for the dimension you are classifying. If we select the "only set unset values" option, it will filter the report to only include Unspecified classification values.
- Based on the values from the first steps, it will run the rules we defined and create a list of new Classification entries.
- That list will then be uploaded and inserted into the lookup table just like any other Classification.
This means that there is no difference in where a Classification came from. We can upload it manually or use Rule Builder, it will always end up in the same lookup table. There is no dependency on the timing of values, besides that new values can override already set values, so the report will always include whatever can be found in the lookup table at the time the report runs.
If you run a report today for today's data based on a Classification, then change the classified value, and then run the same report for today's data tomorrow, the report from tomorrow will show the updated value that would have been reported differently today. The 6-month-limit in the Rule Builder only limits the date range for which it queries for tracked values, not which values are reported.
