The data pipeline functionality enables you to load data in the data format. Use these examples to see how you can access the structure in different ways. For an overview of loading Parquet data, see Data Pipeline Load of Parquet Data from S3.

The SELECT statement for a Parquet file uses the same syntax as the existing JSON selectors. This selection applies to lists, structures, and maps.

The System retrieves its list of requested columns from the SELECT SQL statement. The columns correspond to the field names of the schema for the Parquet files for the load. The load retrieves required columns from the Parquet source files and leaves the other columns alone.

This example shows a list of ["id", "name", "array[]", "a[].project", "my_map.key_name"] as the fields to select from the data. The statement only retrieves the columns for ["id", "name", "array", "a", "my_map"] in the row groups.

JSON Selectors Examples in Data Pipelines

Parquet Selectors Examples in Data Pipelines

Data Pipeline Load of Parquet Data from S3