How to change dtype of array of objects to string

Question

My question is similar to this question How to solve the AttributeError:'list' object has no attribute 'astype'? however, I'm trying to create a series of strings, in order to plot in graph.

 feature_names = df.iloc[:,maxIndex].columns
 newMaxName.append(feature_names)
 newMaxName = np.array(newMaxName, dtype =np.str)
 newMaxName

so it gives me output as

array(['shar', 'dec', 'intel', 'like', 'amb', 'met', 'pid', 'prob_o', 'pf_o_sin', 'pf_o_att', 'pf_o_sha', 'pf_o_fun', 'age_o', 'pf_o_amb', 'pf_o_int', 'attr_o', 'attr', 'attr2_1', 'sinc', 'shar1_1', 'shar_o', 'amb_o', 'intel1_1', 'attr1_1', 'fun_o'], dtype='<U8')

I want them to be string how can I convert them?

Do you want to extract list of strings? .tolist() would do that. These are already strings in UTF-8 — anurag
– anurag, Commented Dec 30, 2020 at 12:26

juanpa.arrivillaga · Accepted Answer · 2020-12-30 12:29:19Z

1

You already did convert it to a string datatype, the default unicode string datatype:

>>> arr
array(['shar', 'dec', 'intel', 'like', 'amb', 'met', 'pid', 'prob_o',
       'pf_o_sin', 'pf_o_att', 'pf_o_sha', 'pf_o_fun', 'age_o',
       'pf_o_amb', 'pf_o_int', 'attr_o', 'attr', 'attr2_1', 'sinc',
       'shar1_1', 'shar_o', 'amb_o', 'intel1_1', 'attr1_1', 'fun_o'],
      dtype='<U8')

If you want byte-strins, be specific:

>>> arr.astype(np.dtype('S'))
array([b'shar', b'dec', b'intel', b'like', b'amb', b'met', b'pid',
       b'prob_o', b'pf_o_sin', b'pf_o_att', b'pf_o_sha', b'pf_o_fun',
       b'age_o', b'pf_o_amb', b'pf_o_int', b'attr_o', b'attr', b'attr2_1',
       b'sinc', b'shar1_1', b'shar_o', b'amb_o', b'intel1_1', b'attr1_1',
       b'fun_o'], dtype='|S8')

Note, byte-strings may be preferable, for memory purposes:

>>> arr.nbytes
800
>>> arr.astype(np.dtype('S')).nbytes
200

However

If your purpose is to plot, just working with object is almost certainly ok. Note, prior to version 1.00 pandas does not use any string-specific types, object was the string type.

If you are working with pandas you may want:

>>> import pandas as pd
>>> pd.__version__
'1.1.1'
>>> s = pd.Series(['abc','def', 'ghi'])
>>> s
0    abc
1    def
2    ghi
dtype: object
>>> s.astype(pd.StringDtype())
0    abc
1    def
2    ghi
dtype: string

answered Dec 30, 2020 at 12:29

juanpa.arrivillaga

97.6k14 gold badges141 silver badges190 bronze badges

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

2 Comments

Razbolt Over a year ago

Yes ı want to plot in graph as yticks, like this ` plt.yticks(y_p,newMaxName) `, in the way arr.astype(np.dtype("S)) it works but the it work in horizontal , but it should be vertical since ı put them in graph, and the last part you said for pd.Series, this strings is the columns of the pandas

juanpa.arrivillaga Over a year ago

@Razbolt it works the same for a column in a dataframe. Your other problem seems totally unrelated to this. Again, almost certainly, you didn't have to do any conversion to begin with.

Collectives™ on Stack Overflow

How to change dtype of array of objects to string

1 Answer 1

However

2 Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

However

2 Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related