python - drop rows with errors for pandas data coercion -

i have dataframe, need convert columns floats , ints, has bad rows, ie., values in column should float or integer instead string values.

if use df.bad.astype(float), error, expected.

if use df.bad.astype(float, errors='coerce'), or pd.to_numeric(df.bad, errors='coerce'), bad values replaced np.nan, according spec , reasonable.

there errors='ignore', option ignores errors , leaves erroring values alone.

but actually, want not ignore errors, drop rows bad values. how can this?

i can ignore errors , type checking, that's not ideal solution, , there might more idiomatic this.

example

test = pd.dataframe(["3", "4", "problem"], columns=["bad"]) test.bad.astype(float) ## valueerror: not convert string float: 'problem' 

i want this:

pd.to_numeric(df.bad, errors='drop') 

and returns dataframe 2 rows.

since bad values replaced np.nan not df.dropna() rid of bad rows now?

edit: since need not drop initial nans, maybe use df.fillna() prior using pd.to_numeric