Pandas 数据帧产生 KeyError: -1

Question

I got the error KeyError: -1 on the following code:

u = []

for i in range(len(df['Some column'])):
   if df['Some column'][i] > df['Some column'][i-1]:
      u.append(df['Some column'][i])

print(u)

This is a pandas dataframe, where I try to print out the list of indexes [i] which are bigger than the previous index [i-1]. But it doesn't work, and I don't know what I am doing wrong.

Answer 1

When you start the loop, i equals 0 , so i-1 equals -1 , which probably isn't in your index.

You could try

u = []

for i in range(1, len(df['Some column'])):
   if df['Some column'][i] > df['Some column'][i-1]:
      u.append(df['Some column'][i])

print(u)

Answer 2

don't use loops you're losing the core functionality of pandas which is to take advantage of vectorised solutions.

we can use shift and .tolist to grab your desired result.

import numpy as np
import pandas as pd

np.random.seed(50)

df = pd.DataFrame({'data' : np.random.randint(0,500,size=500)})

u = df.loc[df['data'] > df['data'].shift(-1)]['data'].tolist()

print(u)
out:
[480, 289, 478, 229, 278, 258, ...]
len(u)
out:
244

Answer 3

The reason might lie in the fact that the index of your data frame may not be perfectly ordered from 0 to range(df). That is, your index may not necessarily be 1, 2, 3, 4,..., N. If your index is something like this:

1, 2, 30, 34, 45, 48, 50

Then, when you run the for loop , you will get this error. You might wanna try this first:

df = df.reset_index()