MongoDB to Silver: Flatten Nested Arrays Before They Double Revenue
A nested `items` array hit gold twice and finance saw 1.9x revenue. Land the document raw, explode once in silver, and do not pretend Mongo is the warehouse.
- mongodb
- modeling
All stacks · Operational data stores
Document sources, change streams, and how they land in the warehouse.
A nested `items` array hit gold twice and finance saw 1.9x revenue. Land the document raw, explode once in silver, and do not pretend Mongo is the warehouse.
MongoDB shows up as a source: document-shaped operational data, change streams, and a schema that “does not exist” until it breaks your warehouse model. The DE job is CDC, flattening, and not copying BSON blobs into gold.
Land raw, yes. Serve raw, no. Flatten to a declared grain in silver. Nested arrays that become facts are how revenue double-counts.
Change streams are native; Debezium has a Mongo connector. Pick the one your platform already operates and can alert on lag.
Rarely. It is a product database. Analytics belong in a columnar store. Dual-purpose Mongo as the warehouse is how you buy a bigger Atlas tier.
One email a week with our latest tutorials. No spam.
Newsletter signup is not live yet. Use the contact form if you want to be notified.