Converting string dictionary to numpy array

Question

I have to change a string dictionary which kind of looks like this.

result = {'top-L': ' ', 'top-M': ' ', 'top-R': ' ',
          'mid-L': ' ', 'mid-M': ' ', 'mid-R': ' ',
          'low-L': ' ', 'low-M': ' ', 'low-R': ' '}

Now, i have to convert this into a 3x3 array.

So, far i have reached a code which do almost what i intended but still its messy.

Here is the code.

import numpy as np
names = ['left','middle']
formats = ['S3','S3']
dtype = dict(names = names, formats=formats)
array = np.fromiter(result.items(), dtype=dtype, count=len(result))
arr = np.reshape(array, (3,3))
print(repr(arr))
print (arr[0][1])

and the output which generated.

array('lo[[(b'top', b' '), (b'top', b' '), (b'top', b' ')],
       [(b'mid', b' '), (b'mid', b' '), (b'mid', b' ')],
       [(b'low', b' '), (b'low', b' '), (bw', b' ')]],
      dtype=[('left', 'S3'), ('middle', 'S3')])
(b'top', b' ')

note print (arr[0][1]) generates (b'top', b' ') which is not expected.

There might be something wrong with this code, any suggestions.

my expected should be like, i can use the dictionary key values using numpy array indexing like arr[0][0] should return "top-L" and [1][1] = ' ', also the output it kind of wrong the keys are getting wrong in the output. — user3814582
– user3814582, Commented May 3, 2018 at 10:33
Your input data seems a bit strange. Are all the values really strings containing a single space? — PM 2Ring
– PM 2Ring, Commented May 3, 2018 at 10:34
You're ever that dictionaries don't preserve the order (all the versions before Python-3.7), right? What's your Python version? — Kasravnd
– Kasravnd, Commented May 3, 2018 at 10:35
@PM2Ring the value of each dict keys will be updated at run time, i order to compare these values and keys, i thought i could use arrays, the computation will be really that way. — user3814582
– user3814582, Commented May 3, 2018 at 10:47

Kasravnd · Accepted Answer · 2018-05-03 10:49:29Z

1

First thing that you need to consider is that dictionaries in lower versions of Python-3.7 do not preserve the order of their items. Therefore if you're using one of these versions you must not expect to get a result with your intended order.

By that being said, generally, a very optimized and handy way to preserve string items in a Numpy array is to use numpy.chararray() objects. As it's also mentioned in documentation chararrays provides a convenient view on arrays of string and unicode values.

Here is how you can get your expected array using chararray :

>>> items = list(result.items())
# By passing `itemsize=5` to the chararray function you're specifying
# the length of each array item
>>> arr = np.chararray((len(items), 2), itemsize=5)
>>> arr[:] = items
>>> arr
chararray([[b'top-L', ''],
           [b'top-M', ''],
           [b'top-R', ''],
           [b'mid-L', ''],
           [b'mid-M', ''],
           [b'mid-R', ''],
           [b'low-L', ''],
           [b'low-M', ''],
           [b'low-R', '']], dtype='|S5')
>>> arr[0]
chararray([b'top-L', ''], dtype='|S5')
>>> arr[0][1]
''

This code has been ran in a Python-3.7 interactive shell environment and that's why the array's order is the same as the dictionary's items order.

answered May 3, 2018 at 10:49

Kasravnd

108k19 gold badges167 silver badges195 bronze badges

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

3 Comments

user3814582 Over a year ago

can it be converted in 3*3 array? Easy to form a loop.

Kasravnd Over a year ago

@user3814582 What kind of 3*3 array? what would be third item?

user3814582 Over a year ago

i am considering [b'top-L', ' '] as one so i have 9 elements so it can be arranged, but then i might not able to access "top-L" and " " values individually, right?

Collectives™ on Stack Overflow

Converting string dictionary to numpy array

1 Answer 1

3 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

3 Comments

Your Answer

Sign up or log in

Post as a guest

Related