Python NumPy Array Data Type

Question

import numpy as np 
d=np.dtype([('name',np.str_),('salary',np.int64)]) 
arr = np.array([('Vedant',80000),('Subodh',4000)],dtype=d) 
print(arr) 

Output:

[('', 80000) ('', 4000)]

Why are the strings blank?

Answer 1

You need to set how many chars you want to insert. ('name',np.str_, 4) this expect each word has only four chars. Or use ('name','U10') for considering ten chars, ('name','U18') for considering 18 chars without specifying.

import numpy as np 
d=np.dtype([('name',np.str_, 4),('salary',np.int64)]) 
# we can define dtype like below without specifying the length of words
# BUT we should pay attention 'U10' only considers ten chars
# OR 'U18' for considering 18 chars
# d=np.dtype([('name','U10'),('salary',np.int64)]) 
arr = np.array([('abcde',80000),('Subodh',4000)],dtype=d) 
print(arr) 
# [('abcd', 80000) ('Subo', 4000)]
# skip 'e' from 'abcde'
# skip 'dh' from 'Subodh'

• hpaulj: I would use the ('name', 'U10') style. The shape for the float is an unnecessary complication here.