Hi! I want to reduce the number of files my parquet sink is producing (one per checkpoint, bucket and parallelism). I remember reading about how I can reshuffle the file sink data to key by bucket assigner, so that each bucket only has one task slot handling it and thus reducing the number of files (at the risk of introducing data skew). Unfortunately either I remember wrong or I am unable to find that particular section, including the configuration setting, in the Flink Documentation 😕 Could anyone help me?