Removing all rows that are not a number in a certain column-CodePudding

I have a problem. I want to remove all rows that are not a number in the column ['a','b']. I have tried this but my approach does not work very well.

Dataframe

        a   b  c
0     0.1  10  x
1     0.5   5  y
2  10 / 5  60  z
3     9.0  12  w
4     125   a  w

Code

import uuid
import pandas as pd
df = pd.DataFrame(
    {'a': [0.1,0.5,'10 / 5', 9.0, 125],
     'b': [10, 5, 60, 12, 'a'],
     'c': ['x', 'y', 'z', 'w', 'w']
    })

print(df)
df['id'] = df.apply(lambda x: uuid.uuid4().int, axis=1)
df_ = df[['a', 'b', 'id']].apply(lambda x: pd.to_numeric(x, errors='coerce')).dropna()
df.merge(df_, left_on=['id'], right_on=['id'], how='inner')

What I want

        a   b  c
0     0.1  10  x
1     0.5   5  y
3     9.0  12  w

CodePudding user response：

You could just do

out = df[~df[['a', 'b', 'id']].apply(lambda x: pd.to_numeric(x, errors='coerce')).isna().any(1)]

CodePudding user response：

what about that

new_df = df[list(df.a.str.isnumeric() != False) and list(df.a.str.isnumeric() != False)]


    a   b   c
0   0.1 10  x
1   0.5 5   y
3   9.0 12  w
4   125 a   w

I am sure there is a nicer version doing this, but isnumeric() doesn't work as expected (by my)