COPY INTO <location> unloads a table or the result of a query to files in an internal stage, an external stage or cloud storage on S3, GCS or Azure. Output is CSV, JSON (VARIANT columns only) or Parquet, with compression options on each. Snowflake splits the output at MAX_FILE_SIZE, 16 MB (16,777,216 bytes) by default and 5 GB (5,368,709,120 bytes) at most. SINGLE = TRUE writes one file instead of many, named data unless a filename and extension are given in the path, and it cannot be combined with INCLUDE_QUERY_ID = TRUE.
What it holds
Warehouse tables and views with their rows, across databases and schemas.
Find the size
Query SNOWFLAKE.ACCOUNT_USAGE.TABLE_STORAGE_METRICS, the per table figures storage billing is calculated on: ACTIVE_BYTES is the data currently queryable, TIME_TRAVEL_BYTES is deleted data still inside Time Travel retention, FAILSAFE_BYTES is data past that but inside Fail-safe, and RETAINED_FOR_CLONE_BYTES is deleted data a clone still references. ACCOUNTADMIN can read it by default; any other role needs GRANT IMPORTED PRIVILEGES ON DATABASE SNOWFLAKE or one of the SNOWFLAKE database roles. That view lags up to 90 minutes, and Account Usage views in general run 45 minutes to 3 hours behind while holding 365 days of history.
What to enter on the form
Snowflake in GB, and the year your company started using it. Seller form