Hi I had similar issues with parquet files when trying to query athena, 

fix was i had to inspect the parquet file since it contained columns such as "Name", "name" which the aws crawler / athena would interpret as a duplicate column since it would see "name" and "name" or "Name" and "Name".