Can I replace empty strings with NA? (opposite of fillna )

[Binger-cn]

I used str.extract_regex() to extract several rows, and using arrow's flatten, and vaex's from_arrays, I composed the result into a DataFrameLocal object.

However, I want to replace the '' values with None, how can I achieve this in a fast way?

Now I can only use .apply(lambda x:None if x=='' else x), which is slow.

update: I found a way: using numpy.place(...). It's faster, but need more memory converting to np.array()

update2: the best way without copying any data, may be using Expression.where():
(df.fieldA=='').where(None, df.fieldA)