Details
-
Sub-task
-
Status: Resolved
-
Major
-
Resolution: Fixed
-
None
-
None
-
None
Description
Is it a norm to fill unmatched columns with nulls as we do ?
protected void tryFillNullInArrayForUnexpectedColumn(Column column, Object[] array, int index) throws SqoopException { if (!column.isNullable()) { throw new SqoopException(SchemaError.SCHEMA_0004, "Target column " + column + " didn't match with any source column and cannot be null."); } LOG.warn("Column " + column + " has no matching source column. Will be ignored."); array[index] = null; }
I looked at kites schema resolution is neat
http://kitesdk.org/docs/0.17.1/Schema-Evolution.html
similar concepts can be applied in sqoop and it relies on of how avro does reader/ writer schema resolution