Protobuf (Value) format

This is the default format. Data is returned row by row as Protobuf: each row is a serialized set of named values with their YQL types. The SDK converts values to native types for your language and exposes a type-safe API.

Use this format when you need:

• Transactional (OLTP) workloads with point reads of individual rows.
• Applications that process small fragments of a table at full row width.
• Native integration with your language’s types.

YQL type mapping

Mapping from YQL types to native language types is implemented in the SDK and depends on the SDK you use. See your SDK’s documentation for the exact mapping.

Returned schema

The schema is a list of columns with their YQL types. That is enough for the SDK to interpret values: given the column type, each value is converted to a native type.

SDK examples

pool = ydb.QuerySessionPool(driver)

query = """
    SELECT * FROM example ORDER BY Key LIMIT 100;
"""

result = pool.execute_with_retries(
    query,
    result_set_format=ydb.QueryResultSetFormat.VALUE,
    schema_inclusion_mode=ydb.QuerySchemaInclusionMode.FIRST_ONLY,
)

for result_set in result:
    print(f"Record batch with {len(result_set.rows)} rows and {len(result_set.columns)} columns")