Changes to the object schema take precedence over changes to the field metadata in the mapping. For example, you add a field to the source object and then edit the metadata of an existing field in the mapping. At run time,

Data Integration

adds the new field and does not edit the existing field.

Data Integration

resolves parameters before picking up the object schema.

Data Integration

treats renamed fields as deleted and added columns. If you rename a field, you might need to update transformations that reference the renamed field. For example, if you rename a field that is used in the lookup condition, the lookup cannot find the new field and the task fails.

When you rename, add, or delete fields, you might need to update the field mapping. For example, if you delete all the previously mapped fields in a target object, you must remap at least one field or the task fails.

Data Integration

writes Null values to a target field in the following situations:

You rename a target field with automatic field mapping, and the field name does not match a source field.
You rename a source field with manual field mapping, and you do not remap the field to the target.

If you delete a field from a source or lookup object and a downstream transformation references the field, the task fails.

If you change a source or lookup field type, the task might fail if the new field type results in errors downstream. For example, if you change an integer field in an arithmetic expression to a string field, the expression is not valid and the task fails.

If you change a target field type,

Data Integration

converts the data from the incoming field to the new target field type. If the conversion results in an error,

Data Integration

drops the row. For example if you change a string type to a date type where the string does not contain a date,

Data Integration

drops the row.