Trying to remove all rows without a numeric value in a column-CodePudding

I have been trying to remove all rows from a dataframe based on the condition that a column doesn't have a numeric value(int, float, etc..)

This is what I tried but it's not working...

city_df = city_df[city_df["city_latitude"].isnumeric()]

Does anyone know how to make it work?

Sample Data:

CodePudding user response：

You can try with a little tweak in the code:

city_df = city_df[city_df['city_latitude'].map(lambda x: str(x)).str.isnumeric().fillna(True)].dropna(subset=['city_latitude'])

Based on this data:

  city_latitude
0      36.35665
1     39.174503
2           NaN
3            AB
4            1C
5     35.961561
6     38.798958

The following result is obtained:

  city_latitude
0      36.35665
1     39.174503
5     35.961561
6     38.798958

EDIT:

Given your comments and errors I suggest a different approach:

city_df['city_latitude'] = pd.to_numeric(city_df['city_latitude'],errors='coerce')
city_df = city_df.dropna(subset=['city_latitude'])

Or in a single line:

city_df = city_df[~pd.to_numeric(city_df['city_latitude'],errors='coerce').isna()]

CodePudding user response：

Try this code

city_df=city_df[city_df['city_latitude'].apply(lambda x: type(x) in [int, np.int64, float, np.float64])]

or
following code will help you to understanding about the work flow

import pandas as pd
import numpy as np
data = {
"city_latitude": [420, 380, 390,'a',5]

}
df = pd.DataFrame(data)
df=df[df['city_latitude'].apply(lambda x: type(x) in [int, np.int64, 
float, np.float64])]

print(df)

output:

  city_latitude
0           420
1           380
2           390
4             5