Direct Load versus Big Data Import

Direct Load versus Big Data Import
Direct Load	Big Data Import
The `Order by` query is used on any attribute or a set of attributes to decide which record can survive in case of duplicate records.	A record is randomly selected and can survive in case of duplicate record detection. (Duplicate records indicate the records that contain same `PRODUCTID` and `EXTENSION`.)
You can choose to save a single version of the record or all versions of the record.	Only one record is retained, which is randomly selected.
Records are processed earlier and then the relationships are processed. If the record is in the same row where relationship is and the record fails; however, the same record is present in some other row and it succeeds, then all relationships are processed.	Records and relationships are processed at the same time. If the record fails and the record is part of another row, relationships are not established. However, if the other row is processed earlier, then relationship is also processed.
In case of hierarchical input maps, you can select additional keys as join keys. This is done because the relationships are processed later.	Supports only a single data source thus, defining additional join keys are not possible.
The Contains field is used to process relationships.	The Contains field is not supported.
Import workflow can also be used to delete records which are not part of import.	Not supported