Replace a specific range of values in a pandas dataframe

Question

I have big data set and there are tons of values which are way over average. For example,

    A         B
1  'H'       10
2  'E'    10000
3  'L'       12
4  'L'        8
5  'O'       11

and I want to set B2 cell as 0 and I tried this,

df['B'] = df['B'].replace([df['B'] > 15], 0)

But didn't get any luck. How can make my data frame like this,

    A         B
1  'H'       10
2  'E'        0
3  'L'       12
4  'L'        8
5  'O'       11

Thank you!

TrebledJ · Accepted Answer · 2019-05-31 09:07:58Z

16

You are really close - instead of replace, use mask:

df['B'] = df['B'].mask(df['B'] > 15, 0)
print (df)
     A   B
1  'H'  10
2  'E'   0
3  'L'  12
4  'L'   8
5  'O'  11

Alternative:

df['B'] = np.where(df['B'] > 15, 0, df['B'])
print (df)
     A   B
1  'H'  10
2  'E'   0
3  'L'  12
4  'L'   8
5  'O'  11

If you want replace some range:

df['B'] = np.where(df['B'].between(8,11), 0, df['B'])
print (df)
     A      B
1  'H'      0
2  'E'  10000
3  'L'     12
4  'L'      0
5  'O'      0

TrebledJ

9,0237 gold badges28 silver badges50 bronze badges

answered Sep 12, 2017 at 5:20

jezrael

868k103 gold badges1.4k silver badges1.3k bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

Watch out with 'np.where'. For some reason it removes the data type of the whole column.

DYZ · Accepted Answer · 2017-09-12 05:23:11Z

6

Another alternative:

df.loc[df['B'] > 15, 'B'] = 0
#   df
#    B
#0  10
#1   0
#2  12
#3   8
#4  11

answered Sep 12, 2017 at 5:23

DYZ

57.3k10 gold badges73 silver badges101 bronze badges

Chokes on DataFrames that have a MultiIndex.