Merging schemas when reading parquet files fails because of incompatible data types int and bigint

Question

When trying to load parquet files with schema merge

   df = spark.read.option("mergeSchema", "true").parquet(\'some_path/partition_date\')