9 new pandas updates that will save you time

[ad_1]

January 25, 2018 · Python tutorial

Since launching my Python pandas video series in 2016, there have been 10 new releases of the pandas library, together with tons of of latest options, bug fixes, and API modifications. I used to be operating model 0.18 on the time, however I’ve lastly upgraded to model 0.22 (the newest launch).

Every time I replace a Python library, I wish to know two issues:

  1. What are the new options I’d wish to use?
  2. What are the API modifications I have to learn about?

To determine this out, I spent a number of hours skimming by means of the pandas release notes, compiling an inventory of modifications that could be of curiosity to me and my college students. Then I recorded two new pandas movies explaining these modifications, so as to profit from my analysis!

  1. Create a datetime column from a DataFrame (1:18)
  2. Create a class column throughout file studying (4:24)
  3. Convert the information sort of a number of columns without delay (7:45)
  4. Apply a number of aggregations on a Sequence or DataFrame (9:48)
  5. Bonus: Obtain the official pandas cheat sheet (13:14)

Notice: Click on on a timestamp to leap on to that time within the video, or open the Jupyter notebook for those who simply wish to learn the code.

  1. ix has been deprecated (1:19)
  2. Aliases have been added for isnull and notnull (7:04)
  3. drop now accepts “index” and “columns” key phrases (10:37)
  4. rename and reindex now settle for “axis” key phrase (13:35)
  5. Ordered classes have to be specified unbiased of the information (16:18)

Notice: Click on on a timestamp to leap on to that time within the video, or open the Jupyter notebook for those who simply wish to learn the code.

In case you’re comparatively new to pandas, these two movies won’t make loads of sense, so I might recommend beginning with my complete pandas video series as an alternative.

In order for you an extended introduction to any of the matters coated above, take a look at these associated movies:

If it is advisable to determine what model of pandas you are operating, merely import pandas as pd, then run pd.__version__ (observe the 2 double underscores).

If you wish to replace pandas, you should utilize conda for those who’re operating the Anaconda distribution, or pip in any other case.

Thanks for watching! When you have any questions, I would love to listen to from you within the feedback part under.

P.S. You may additionally be fascinated about my latest pandas tutorial, Data science best practices with pandas.

[ad_2]

Source link

Write a comment