The output value is always a single record with no header in the CSV data. Each output DATASET value includes a schema if any of the following occurs:
- The VARCHAR/CLOB CSV file has a header record
- The input value is a DATASET value with a schema
- The USING SCHEMA clause specifies a schema
The output DATASET values do not include the header record from the input file, but the field_names key places the header names in the schema.
Use CSVSplit to populate a destination table with output DATASET values. Because the values adhere to the same schema, use a schema associated with the destination table column by WITH SCHEMA so the schema is not stored in every row. The schema included with the output DATASET value is discarded if a schema is already defined for the table column. The table column schema must match the schema in the output DATASET value.
CSVSplit uses the following clauses to compose DATASET data type instances:
- SCHEMA
- RETURNS