The pandas I/O System

So far, we have been creating our pd.Series and pd.DataFrame objects inline with data. While this is helpful for establishing a theoretical foundation, very rarely would a user do this in production code. Instead, users would use the pandas I/O functions to read/write data from/to various formats.

I/O, which is short for input/output, generally refers to the process of reading from and writing to common data formats like CSV, Microsoft Excel, JSON, etc. There is, of course, not just one format for data storage, and many of these options represent trade-offs between performance, storage size, third-party integration, accessibility, and/or ubiquity. Some formats assume well-structured, stringently defined data (SQL being arguably the most extreme), whereas other formats can be used to represent semi-structured data that is not restricted to being two-dimensional (JSON being great example).

The fact that pandas can interact with so many of these data formats...