You need to follow certain rules and guidelines to configure tasks that uses Open Table
Connector.
Consider the following rules and guidelines when you configure a task:
A mapping fails at run time when you
specify a simple filter with
Starts With
,
Ends
With
, or
Contains
operators to retrieve
data. To avoid this issue, configure an advanced filter with a
like
query to filter query results based on patterns
instead of exact textual matches.
When you configure an advanced data
filter for a field of Timestamp data type, the filter ignores the time and
considers only the date.
When you use a simple filter on the date or timestamp data type column to
read data, the task fails at run time with the following error:
An error occurred during mapping translation: Invalid expression
string for filter condition
As a workaround, use the advance filter to filter date and timestamp data
type columns from an Open Table source.
You can't use uncached lookup in the
Lookup transformation.
When you run a mapping with an Open Table
connection to read from or write data to Apache Iceberg tables available in AWS
Glue Catalog, the mapping fails if the source and target are pointing to
different AWS Glue catalogs available in different
regions.
When you run a mapping with an
Upsert target operation and all the condition fields are not mapped in the field
mapping, the mapping fails.
When you run a mapping to read
from a Snowflake table and write to an Apache Iceberg table managed by the Hive
Metastore catalog, the mapping fails.
When you configure a mapping with an Apache
Iceberg table, the mapping fails in the following scenarios:
If the source table is
managed by the AWS Glue Catalog and the target table is managed by the
Hive Metastore catalog.
If the source table is
managed by the Hive Metastore catalog and the target table is managed by
the AWS Glue Catalog.
When you configure a filter query
in a mapping with an AND or OR operator to read data from an Apache Iceberg
table managed by the Hive Metastore catalog, the query does not return any
records.
When you read data from an Apache
Iceberg table managed by the Hive Metastore catalog in a mapping, the data
preview at the source table and the preview job fail.