A mapping configured to read from Google BigQuery source and write to a Parquet file in Google Cloud Storage target fails in the following cases:

Data is of the Date data type and the date is less than 1582-10-15.
Data is of the Int96 data type and the timestamp is less than 1900-01-01T00:00:00Z.

spark.sql.parquet.int96RebaseModeInWrite=LEGACY
spark.sql.parquet.datetimeRebaseModeInWrite=LEGACY
spark.sql.parquet.int96RebaseModeInRead=LEGACY
spark.sql.parquet.datetimeRebaseModeInRead=LEGACY
spark.sql.avro.datetimeRebaseModeInWrite=LEGACY
spark.sql.avro.datetimeRebaseModeInRead=LEGACY

When you run a mapping in advanced mode to read from or write to fields of the Date and Timestamp data types, the time zone defaults to the Secure Agent host machine time zone.

To change the Date and Timestamp to the UTC time zone, you can either set the Spark properties globally in the Secure Agent directory for all the tasks in the organization that use this Secure Agent, or you can set the Spark session properties for a specific task from the task properties:  

To set the properties globally, perform the following tasks:

Add the following properties to the
<Secure Agent installation directory>/apps/At_Scale_Server/41.0.2.1/spark/custom.properties
directory:
infacco.job.spark.driver.extraJavaOptions=-Duser.timezone=UTC
infacco.job.spark.executor.extraJavaOptions=-Duser.timezone=UTC
Restart the Secure Agent.

To set the properties for a specific task, navigate to the Spark session properties in the task properties, and perform the following steps: