How to access multiple fields from a structured numpy.array?

问题

I came across this difficulty accessing multiple fields (columns)

input:

a = np.array([(1.0, 2,1),(3.0, 4,2),(9, 3,6)], dtype=[('x', float), ('y', float), ('z', float)])
a=np.reshape(a,(a.shape[0],-1))
a

output:

array([[(1.0, 2.0, 1.0)],
       [(3.0, 4.0, 2.0)],
       [(9.0, 3.0, 6.0)]], 
      dtype=[('x', '<f8'), ('y', '<f8'), ('z', '<f8')])

if i want to access the first column i can do:

in: a[:]['x']

out: array([[ 1.], [ 3.], [ 9.]])

but what is the right syntax if i want to access (for example) first an 3rd column? Something like

in: a[:]['x':'z']

obviously does not work

回答1:

Use a list of field names as an index to the array. The result is an array of the same shape, but with only the selected fields in the records (array elements with multiple fields are called records).

import numpy as np
a = np.array([(1.0, 2,1),(3.0, 4,2),(9, 3,6)], dtype=[('x', float), ('y', float), ('z', float)])
print(a)

print(a[['x', 'z']])

You can apply a further level of indexing to the resulting array to select only the required elements, should you choose.

回答2:

a[:][['x', 'z']]

Out[9]: 
array([[(1.0, 1.0)],
   [(3.0, 2.0)],
   [(9.0, 6.0)]],

Pass the column names as a list

回答3:

Consider having a lot of columns and you do not want to add all items manually - you can do: The column names of a are converted to a list. Afterwards you can access the columns as a list indexwise or itemwise

col_to_ex=list(a.dtype.names)
col_to_ex=col_to_ex[0]+...

col_to_ex=list(a.dtype.names).remove('y')

and then you can do:

a[:][col_to_ex]

来源：https://stackoverflow.com/questions/24271828/how-to-access-multiple-fields-from-a-structured-numpy-array

标签

python

numpy

field