Pandas API on Spark

• Options and settings
  • Getting and setting options
  • Operations on different DataFrames
  • Default Index type
  • Available options
• From/to pandas and PySpark DataFrames
  • pandas
  • PySpark
• Transform and apply a function
  • transform and apply
  • pandas_on_spark.transform_batch and pandas_on_spark.apply_batch
• Type Support in Pandas API on Spark
  • Type casting between PySpark and pandas API on Spark
  • Type casting between pandas and pandas API on Spark
  • Internal type mapping
• Type Hints in Pandas API on Spark
  • pandas-on-Spark DataFrame and Pandas DataFrame
  • Type Hinting with Names
  • Type Hinting with Index
• From/to other DBMSes
  • Reading and writing DataFrames
• Best Practices
  • Leverage PySpark APIs
  • Check execution plans
  • Use checkpoint
  • Avoid shuffling
  • Avoid computation on single partition
  • Avoid reserved column names
  • Do not use duplicated column names
  • Specify the index column in conversion from Spark DataFrame to pandas-on-Spark DataFrame
  • Use distributed or distributed-sequence default index
  • Handling index misalignment with distributed-sequence
  • Reduce the operations on different DataFrame/Series
  • Use pandas API on Spark directly whenever possible
• Supported pandas API
  • CategoricalIndex API
  • DataFrame API
  • DatetimeIndex API
  • Index API
  • MultiIndex API
  • Series API
  • TimedeltaIndex API
  • General Function API
  • Expanding API
  • ExpandingGroupby API
  • Rolling API
  • RollingGroupby API
  • Window API
  • DataFrameGroupBy API
  • GroupBy API
  • SeriesGroupBy API
• FAQ
  • Should I use PySpark’s DataFrame API or pandas API on Spark?
  • Does pandas API on Spark support Structured Streaming?
  • How is pandas API on Spark different from Dask?