Loading for Snowflakes
Loading is how data enters Snowflake. Files in cloud storage are referenced through stages, parsed with file formats and copied into tables with COPY INTO, while Snowpipe loads new files continuously. The reverse operation unloads query results back to files. Choosing between bulk and continuous loading depends on latency needs and file arrival patterns.
Stage: Data Loading · Part of the Snowflake roadmap · Last reviewed 19 November 2026
What to learn in Loading
- Create an external stage pointing to cloud storage and list its files
- Define a file format for CSV or JSON and load data with COPY INTO
- Handle bad rows using the ON_ERROR option and validate before loading
- Set up Snowpipe for continuous loading of newly arrived files
- Unload query results to a stage as Parquet or CSV
Loading interview questions
What is the difference between COPY INTO and Snowpipe?
COPY INTO is a bulk command you run, using your chosen warehouse. Snowpipe loads files automatically as they arrive in a stage, using Snowflake-managed compute, which suits continuous small batches. COPY INTO fits scheduled larger loads.
How does Snowflake avoid loading the same file twice?
It keeps load metadata per table for a limited period, so COPY INTO skips files it has already loaded unless you set FORCE. Renaming or modifying a file after the metadata expires can cause it to be treated as new.
Where Loading fits in the Snowflake roadmap
Loading belongs to the Data Loading stage (intermediate level). Get data in and out.
Other units in this stage
Back to the full Snowflake roadmap