Throwing away 3 out of 4 samples from a data frame

Question

In Python, I have a DataFrame that looks like the following, all the way down to about 5000 samples:

I was wondering, in pandas, how do I remove 3 out of every 4 data points in my DataFrame?

How do you choose which 3 out of every 4 to remove? – SethMMorton Jul 27 '17 at 20:15 — SethMMorton, Jul 27 '17 at 20:15
It doesn't matter which, for my use case. – Gary Jul 27 '17 at 20:21 — Gary, Jul 27 '17 at 20:21

unutbu · Accepted Answer · 2017-07-27T20:47:12.110

4

To obtain a random sample of a quarter of your DataFrame, you could use

test4.sample(frac=0.25)

or, to specify the exact number of rows

test4.sample(n=1250))

If your purpose is to build training, validation, and testing data sets, then see this question.

edited Jul 27 '17 at 20:47

answered Jul 27 '17 at 20:23

unutbu

It's not related to machine learning, but rather, related to this question: https://stackoverflow.com/questions/45337886/valueerror-cannot-copy-sequence-to-array-axis-in-python-for-matplotlib-animatio/45342706#45342706 – Gary Jul 27 '17 at 20:49

score 1 · Answer 2 · answered Jul 27 '17 at 20:14

1

If you want to select every 4th point, then you can do the following. This will select rows 0, 4, 8, ...:

test4.iloc[::4, :]['Accel']

answered Jul 27 '17 at 20:14

nanojohn

2 Answers2