Why you should not dump DataFrames to a CSV

The CSV file format is widely used to save Pandas DataFrames. But are you aware of its limitations? To name a few,

1. The CSV does not store the datatype information. Thus, if you modify the datatype of column(s), save it to a CSV, and load again, Pandas will not return the same datatypes.

2. Saving the DataFrame to a CSV file format isn't as optimized as other supported formats by Pandas. These include Parquet, Pickle, etc.

Of course, if you need to view your data outside Python (Excel, for instance), you are bound to use a CSV. But if not, prefer other file formats.

---

Further reading: Why I Stopped Dumping DataFrames to a CSV and Why You Should Too.