EDIT 2:
I was able to figure it out. Since the original .parquet "file" is a directory, just loop through the directory with glob for all the inner .parquet files - the original code for Python-parquet works in there.
for filename in glob.glob("/mnt/mapr/saw/user/mapr/{0}.parquet/*.parquet".format(tempTableID)): with open(filename) as foo: for row in parquet.DictReader(foo, columns=["column"]): print(json.dumps(row))