Examples
Take separate series and convert to numeric, coercing when told to
import numpy as np
import pandas as pd
s = pd.Series(['2.0', '4', -2])
pd.to_numeric(s)
pd.to_numeric(s, downcast='float')
pd.to_numeric(s, downcast='signed')
s = pd.Series(['mango', '2.0', '4', -2])
pd.to_numeric(s, errors='ignore')
pd.to_numeric(s, errors='coerce')