atoti.AutoMultiColumnArrayConversion#

final class atoti.AutoMultiColumnArrayConversion#

Pass it to a DirectQuery *ConnectionConfig class to automatically convert all external tables with array values stored with one element per column to tables with array columns.

Example

>>> import os
>>> from atoti_directquery_snowflake import ConnectionConfig, TableConfig
>>> url = f"jdbc:snowflake://{os.environ['SNOWFLAKE_ACCOUNT_IDENTIFIER']}.snowflakecomputing.com/?user={os.environ['SNOWFLAKE_USERNAME']}&database=TEST_RESOURCES&schema=TESTS"
>>> connection_config = ConnectionConfig(
...     url=url,
...     auto_multi_column_array_conversion=tt.AutoMultiColumnArrayConversion(
...         separator="_",
...         threshold=3,
...     ),
...     password=os.environ["SNOWFLAKE_PASSWORD"],
... )
>>> external_database = session.connect_to_external_database(connection_config)
>>> external_table = external_database.tables["TUTORIAL", "MULTI_COLUMN_QUANTITY"]

In the external database, external_table has this content:

PRODUCT	QUANTITY_0	QUANTITY_1	QUANTITY_2	QUANTITY_3	QUANTITY_4
product_1	10.0	20.0	15.0	25.0	10.0
product_2	50.0	65.0	55.0	30.0	80.0

It has 5 QUANTITY{separator}{index} columns (where separator is "_"). Since 5 is greater than the threshold passed above, the automatic conversion will activate and these 5 columns will be merged into an array column:

>>> list(external_table)
['PRODUCT', 'QUANTITY']
>>> table = session.add_external_table(
...     external_table,
...     table_name="Sales (Multi column array)",
...     config=TableConfig(keys=["PRODUCT"]),
... )
>>> table.head().sort_index()
                                 QUANTITY
PRODUCT
product_1  [10.0, 20.0, 15.0, 25.0, 10.0]
product_2  [50.0, 65.0, 55.0, 30.0, 80.0]