I daily generate and save a dataset partitioned by the field ‘Hotel’ and I was wondering to know if reading a partitioned file is slower that reading the same file without partitions.

Lets, say that my main first parquet file is (without partitioning):

• main_bookings.parquet
  Then I also have, as I said, a file with its. For example, one partition is:
• main_bookings_partitioned.parquet/Hotel=Barcelona_1
• main_bookings_partitioned.parquet/Hotel=Barcelona_2
• main_bookings_partitioned.parquet/Hotel=Barcelona_3
• main_bookings_partitioned.parquet/Hotel=Madrid_1
• main_bookings_partitioned.parquet/Hotel=…

I was wondering to know if, from Databricks and using PySpark, reading the main_bookings.parquet (without partitions) is much faster than reading main_bookings_partitioned (the same file but with partitions)

Thank you in advance,

I would like to get a response of the solution and understand why this is happening.