How to handle mixed data types in numpy arrays

Question

Stuck in this Numpy Problem

country=['India','USA']
gdp=[22,33]

import numpy as np
a=np.column_stack((country,gdp))

array([['India', '22'],
       ['USA', '33']], dtype='<U11')

I have an NDArray and I want to find the maximum of the 2nd column. I tried the below

print(a.max(axis=1)[1])
print(a[:,1].max())

It threw this error: TypeError: cannot perform reduce with flexible type

Tried converting the type

datatype=([('country',np.str_,64),('gross',np.float32)])

new=np.array(a,dtype=datatype)

But got the below error

could not convert string to float: 'India'.

Somebody edited your question please accept that. So, that we can review your question in more readable form. — SalGorithm
– SalGorithm, Commented Mar 19, 2018 at 9:54
Now wait and keep an eye on your question, you'll be getting your answer as we speak. — SalGorithm
– SalGorithm, Commented Mar 19, 2018 at 10:05

swathis · Accepted Answer · 2018-03-19 10:16:18Z

2

The error is due to the string data in your array, which makes the dtype to be Unicode(indicated by U11 i.e., 11-character unicode) string. If you wish to store data in the numerical format, then use structured arrays. However, if you only wish to compute the maximum of the numerical column, use

print(a[:, 1].astype(np.int).max())
// 33

You may choose to use other numerical dtypes such as np.float inplace of np.int based on the nature of data in the specific column.

edited Mar 19, 2018 at 10:16

answered Mar 19, 2018 at 10:00

swathis

3666 silver badges17 bronze badges

Sign up to request clarification or add additional context in comments.

4 Comments

Priyash Maini Over a year ago

But why did it create a unicode when my list is already float ? but in the array it gets converted to string ? WHy ?

swathis Over a year ago

Firstly, your list is not "float". If you create a numpy array from the list [22,33], the type will be int (as per the data values). Secondly, since you create the array from multiple lists, some of which consist of strings, the type of the resultant array will be string - in order to accommodate the different non-matching types.

Priyash Maini Over a year ago

Secondly, since you create the array from multiple lists, some of which consist of strings, the type of the resultant array will be string - in order to accommodate the different non-matching types ...... How can we avoid such situation when creating array from multiple lists.

swathis Over a year ago

Structured arrays - docs.scipy.org/doc/numpy-1.14.0/user/basics.rec.html address this.

jpp · Accepted Answer · 2018-03-19 10:18:35Z

0

Consider using numpy structured arrays for mixed types. You will have no issues if you explicitly set data types.

This is often necessary, and certainly advisable, with numpy.

import numpy as np

country = ['India','USA','UK']
gdp = [22,33,4]

a = np.array(list(zip(country, gdp)),
             dtype=[('Country', '|S11'), ('Number', '<i8')])

res_asc = np.sort(a, order='Number')

# array([(b'UK', 4), (b'India', 22), (b'USA', 33)], 
#       dtype=[('Country', 'S11'), ('Number', '<i8')])

res_desc = np.sort(a, order='Number')[::-1]

# array([(b'USA', 33), (b'India', 22), (b'UK', 4)], 
#       dtype=[('Country', 'S11'), ('Number', '<i8')])

edited Mar 19, 2018 at 10:18

answered Mar 19, 2018 at 10:12

jpp

166k37 gold badges301 silver badges362 bronze badges

1 Comment

Priyash Maini Over a year ago

But I need a 2D array. these are two different output [('India', 22.) ('USA', 33.)] [['India' '22'] ['USA' '33']]

Collectives™ on Stack Overflow

How to handle mixed data types in numpy arrays

2 Answers 2

4 Comments

1 Comment

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

4 Comments

1 Comment

Your Answer

Sign up or log in

Post as a guest

Linked

Related