numpy data slicing with mixed type

Question

I use numpy.genfromtxt to load data from a csv file, and this file has mixed data type, like the following:

date,value1,value2
1997-02, 432, 1
1997-03, 300, 1
1997-04, 432, 0

I use the following command to load data:

data = numpy.genfromtxt('data/test.csv', dtype=None, delimiter=',', skip_header=1)

and I get

array([(b'1997-02', 432, 1), (b'1997-03', 300, 1), (b'1997-04', 432, 0)], 
      dtype=[('f0', 'S7'), ('f1', '<i8'), ('f2', '<i8')])

Now I want only the date of this data, how can I slice the numpy array with mixed type? I tried to use data[:,0] but it does not work

Answer 1

You can use the field name:

data['f0']

# array([b'1997-02', b'1997-03', b'1997-04'], 
#       dtype='|S7')

---

And further for your reading purpose, I think you want to specify the names = True instead of skip_header so that the first line will be read in as the field names of the structured array:

data = np.genfromtxt('data/test.csv', dtype=None, delimiter=',', names = True)
​    
data
# array([(b'1997-02', 432, 1), (b'1997-03', 300, 1), (b'1997-04', 432, 0)], 
#       dtype=[('date', 'S7'), ('value1', '<i8'), ('value2', '<i8')])

Now you can access date as:

data['date']
# array([b'1997-02', b'1997-03', b'1997-04'], 
#       dtype='|S7')