Ingest Data into Dynamics 365 Customer Insights - Data
Ingestion fundamentals
Ingestion moves raw data into Customer Insights - Data so it can be unified. Candidates must know the supported connector types: Power Query connectors for many SaaS and file sources, Azure Data Lake Storage, Microsoft Dataverse, and direct upload of CSV files. Each data source produces tables; each table has attributes with data types, and ingestion can be scheduled or refreshed manually.
Key concepts
Understand the data source versus table hierarchy: one source can yield many tables. Know that schemas are inferred from the source and can be edited (rename, change type, set primary key). Recognize that incremental refresh reduces load time by bringing only changed rows, and that a full refresh reloads everything. Know how data profiling surfaces quality issues such as nulls, duplicates, and type mismatches before unification. Understand that ingestion errors are surfaced per table and must be resolved before downstream steps run.
Common traps
A frequent trap is confusing the primary key of a source table with the customer ID used in unification-they are different settings. Another is assuming ingestion automatically deduplicates records; it does not. Questions may ask where to fix a wrong data type: in the table's schema/attribute settings, not in unification. Also remember that deleting a data source removes its tables and can break dependent processes.
Worked example
You load a CSV of contacts and a Dataverse table of orders. Create two data sources, confirm each table's attributes and types, set the contact table's primary key to the source contact ID, and verify profiling shows no null keys. Only then proceed to unification, where you choose match columns and order.
Rule of thumb
Validate schema and keys at ingestion; never carry dirty keys into unification.