Modify Column With a Condition

Question

I’m trying to modify a column in a dataframe with a condition. I have two columns on it, named ‘Population’ and ‘Census’. I have got millions of data on it and I just need modify those that ‘Population < Census’. Here is the dataframe very reduced.

dataframe1 = pd.DataFrame()

Population = [34,2,1,4542,422,122,55,7876,45,23,1,3]
Census = [12,43,14,6545,63,123,654,33,123,12,99,88]

it’s placed randomly but it serves to explain.

I want to modify the column ‘Census’. Here what I tried.

a['Census'] = np.where[a['Population']<a['Census'], a['Population'], 1]

But I get an TypeError: ‘function’ object is not subscriptable.

How I can modify this in the same column?

Asked By: DatBigD

||

Source

Answer 1

After np.where you should use ‘(‘ not ‘[‘. This is a function that’s why it is not subscriptable as mentioned in the error. Can you check this one ?

dataframe1["new"] =  np.where(dataframe1['population']<dataframe1["census"], dataframe1["population"], 1)

Answered By: Ugur Yigit

Answer 2

Indeed, np.where is a function so you have to use np.where(...). Alternatively, you can also use DataFrame.where as replacement of Numpy.where:

a['Census'] = a['Census'].where(a['Population'] < a['Census'], other=1)
print(a)

# Output
    Population  Census
0           34       1
1            2      43
2            1      14
3         4542    6545
4          422       1
5          122     123
6           55     654
7         7876       1
8           45     123
9           23       1
10           1      99
11           3      88

Input data:

a = pd.DataFrame({'Population': Population, 'Census': Census})
print(a)

# Output
    Population  Census
0           34      12
1            2      43
2            1      14
3         4542    6545
4          422      63
5          122     123
6           55     654
7         7876      33
8           45     123
9           23      12
10           1      99
11           3      88

Answered By: Corralien

Modify Column With a Condition

Question:

Answers: