pd.read_csv gives me str but need float

Question

I have a CSV which looks like this:

Date,Open,High,Low,Close,Adj Close,Volume
2007-07-25,4.929000,4.946000,4.896000,4.904000,4.904000,0
2007-07-26,4.863000,4.867000,4.759000,4.777000,4.777000,0
2007-07-27,4.741000,4.818000,4.741000,4.788000,4.788000,0
2007-07-30,4.763000,4.810000,4.763000,4.804000,4.804000,0

after

data = pd.read_csv(file, index_col='Date').drop(['Open','Close','Adj Close','Volume'], axis=1)

i end up with a df which looks like this:

                High       Low
Date                          
2007-07-25  4.946000  4.896000
2007-07-26  4.867000  4.759000
2007-07-27  4.818000  4.741000
2007-07-30  4.810000  4.763000
2007-07-31  4.843000  4.769000

Now i want to get High - Low. Tried:

np.diff(data.values, axis=1)

but getting an error: unsupported operand type(s) for -: 'str' and 'str'

but sure why the values in the df are str in the first place. Grateful for any solution.

Answer 1

I think you need to_numeric with errors='coerce' because it seems there are some bad data:

data = pd.read_csv(file, index_col='Date', usecols=['High','Low'])

data = data.apply(pd.to_numeric, errors='coerce')

Answer 2

The read_csv dtype option doesn't work ?

from the documentation dtype : Type name or dict of column -> type, default None Data type for data or columns. E.g. {‘a’: np.float64, ‘b’: np.int32} Use str or object to preserve and not interpret dtype. If converters are specified, they will be applied INSTEAD of dtype conversion.

data = pd.read_csv(file,
    index_col='Date',
    usecols=['High','Low'],
    dtype={'High': np.float64, 'Low': np.float64})