A data source is a container that holds the data you upload to Analytics, and comprises the CSV file you upload and a mapping of existing Analytics fields to the fields in your CSV. Data sources control how uploaded data gets joined with existing data. You configure data sources at the property level.
To manage the data sources for a property, click Admin, then under Data collection and modification, click Data import.
A data-source type corresponds to the specific type of data you want to import. For example, there are data-source types for cost data, item data, user data, and offline-event data. For each one, you have different options for the dimensions and metrics (the schema) you can upload data for.
When you create a source, you define a schema, the structure that joins the data you upload with the existing Analytics event data. A simple schema consists of a join-key dimension (the "key") and the import dimension(s) or metric(s). To import data, Analytics looks for key values in event data that match key values in the uploaded data. When Analytics finds a match, it adds (or replaces if data was already collected) the additional dimension and metric values associated with that key to the existing event data. Some data-source types let you use multiple dimensions to define the key, and most can use multiple dimensions/metrics for the import fields.
The join key is composed of at least one Analytics dimension, and can be composed of up to three.
The key dimensions vary based on the type of data set you choose.
A list of available key dimensions is provided in the drop-down menu in the data-source schema builder.
You can upload data via the same data source multiple times as long as you are just adding values to existing fields. If you want to add fields to a data source, you need to delete the existing data source and then create a new one -- once you save a data source, you can't change the mapping of Analytics fields to the fields in your CSV.