[SOLVED] Discard rows with a string value from a dataframe

Discard rows with a string value from a dataframe

I have a dataframe like this:

C1	C2	C3	C4
1	foo	asd	23
foo	foo	asd	43
3	foo	asd	1
4	foo	asd	bar

I'm trying to filter (and discard) all rows that have strings in C1 or C4 columns, my final dataframe must be:

C1	C2	C3	C4
1	foo	asd	23
3	foo	asd	1

I'm trying to do this using "isNaN" but I'm not sure how I should use it.

This is my code:

df = pd.read_csv(
    path_file,
    sep=",",
    usecols=columns,
    skiprows=0,
    skipfooter=0,
    engine="python",
    encoding="utf-8",
    skipinitialspace=True,
    on_bad_lines='warn',
    names=columns)

df_new = df[df["C1"].notna()]
df_new_2 = df[df["C4"].notna()]

Any idea about how I can achieve this?

Solution

You can try something like this:

df.loc[df[['C1', 'C4']].apply(pd.to_numeric, errors='coerce').dropna(how='any').index]

Output:

  C1   C2   C3  C4
0  1  foo  asd  23
2  3  foo  asd   1

Use pd.to_numeric with errors='coerce' to give NaN that can be used to with dropna.