Is it better to store Pandas data frames in a dictionary or in a panel?

Question

Hope this doesn't sound like an open discussion question. I am going to give some details for my particular case.

I'm new to Pandas, and I need to save some 2D arrays where the columns represent frequencies and rows representing directions (spectra of 2D waves, if you're interested). Each array represents a specific time.

I store these arrays as Pandas DataFrames, but for storing them in one object, I thought of two options:

• Saving DataFrames in a dictionary, where the key is a timestamp.

• Saving DataFrames in the Pandas panel, where the item is a timestamp.

The first option seems simple and has the flexibility to store arrays with different sizes, indexes, and column names. The second option seems to be better for data processing, since the panels have special methods, and can also be easily saved or exported (for example, in csv or pickle).

Which of the two options is best suited for: speed, memory usage, flexibility, and data analysis?

Hello

Answer 1

I don’t think you need a panel. I recommend the nested data approach.