Pandas is finally version 1.0
Can you believe Pandas 1.0.0 is finally here?
⭐️ My top 5 picks for some exciting new features:
1️⃣ New “string” and “bool” data types.
2️⃣ pd.NA: to be used as nullable integer, bool and new string data type (I’m so excited for this one, hopefully makes null handling a lot cleaner). Note that np.nan is still used for float types.
3️⃣ df.to_markdown(): converts dataframe to markdown syntax. I use my own custom function for this so often for documentation, glad to retire my own extension! (also has df.to_latex and .to_json)
4️⃣ Allow the user to choose Numba in rolling.apply instead of Cython.
5️⃣ Defining custom windows for rolling oeprations: this can definitely help clean up all these time series feature engineering scripts.
💎 Bonus: New ignore_index keyword to reset index after sorting or dropping. I mean, it’s a tiny thing but this makes me heart sing.