Partition Parquet File by Date

The pyspark script below can split one single big parquet file into small parquet files based on date column.

from pyspark.sql.functions import col, dayofmonth, month, year

df = #load data frame

dt = col("date").cast("date")
fname = [(year, "year"), (month, "month"), (dayofmonth, "day")]
exprs = [col("*")] + [f(dt).alias(name) for f, name in fname]

(*exprs).write.partitionBy(*(name for _, name in fname)).mode("overwrite").parquet('output_location'))

Leave a Reply

Your email address will not be published. Required fields are marked *