从某个范围内的列中找到一个固定值，并在熊猫数据框中找到另一列的每个唯一值

Question

I have a data frame like this:

df
col1      col2
 1        50000
 1        2000
 2        51000
 3        100
 3        5000
 3        50500
 4        200
 4        51500
 5        49000

I want to identify the values with plus minus 10 percent for each of col1 values which occurs for every col1 unique values.

the final output should look like

col1        col2
  1         50000
  2         51000
  3         50500
  4         51500
  5         49000

if other values other than the values around 50000 presents and have within plus minus 10 percent range, add those with the values around 50000

How to do it using pandas/python with most efficient way ?

Answer 1

Use list cpmprehension for loop by all unique values of col2 , filter by +-10% with Series.between and boolean indexing and compare if all values exist in all groups with set created by col1 . Last filter by Series.isin :

s = set(df['col1'])
print (s)
{1, 2, 3, 4, 5}

a = [x for x in df['col2'].unique() 
     if set(df.loc[df['col2'].between(x - x *.1, x + x*.1), 'col1']) == s]
print (a)
[50000, 51000, 50500, 51500, 49000]

df = df[df['col2'].isin(a)]
print (df)
   col1   col2
0     1  50000
2     2  51000
5     3  50500
7     4  51500
8     5  49000