0

Being a regular user of numpy, I am falling in love with Pandas for its syntax and concepts, and started to read about PyTables, which seems very robust and performance-oriented.

My main doubts are:

  • If I were to use a home-made framework composed by Pandas + PyTables, which layer of the data-processing pipeline (data importing, converting, analyzing, storing, retrieving) should be delegated to each one?
  • What's the natural talent of Pandas and how it is related to the natural talent of PyTables, and how their talents relate?
  • Are they orthogonal/complimentary/independent, or do they overlap functionality somehow?
  • How do they diferentiate from one another?

Thanks for any light!

heltonbiker
  • 26,657
  • 28
  • 137
  • 252

1 Answers1

2

There's not much overlap between pandas and pytables, but they are very good compliments of each other. PyTables is all about storage and retrieval while pandas is about working with the data after retrieval and before storage. Pandas has convenient interfaces to PyTables (Check out pandas.io.pytables.HDFStore) so you can easily store pandas Series/DataFrame using pytables.

Chang She
  • 16,692
  • 8
  • 40
  • 25