Pandas Dataframe Tips

When dropping rows with NaN's in a Pandas Dataframe, this

df = df.dopna()

may be faster than this

df.dropna(inplace=True).

---

Casting converting a column of a dataframe to float (for example) can be very slow if done like this:

df[col1] = pd.to_numericdf(df[col1])

or

df[col1] = df[col1].apply(pd.to_numeric)

or

df[col1] = df[col1].astype(float)

If you know how to do this efficiently, please tell me. As of now, my only work around is to avoid doing this operation if possible.