Python NumPy, can't create array of items array-CodePudding

I have a NumPy dataset with sentence IDs and descriptions values. I would like to create an ordered list in my desired format which is... [['value'], ['value']] containing just description values. The order of the new list must stay the same as the original so I can match them back to the IDs later.

My issue is that I can't achieve the desired format for the new list instead of [['value'], ['value']] I get ['value', 'value'], why?

This is what I have tried:

Original input data:

[
[UUID('11ea1bca-eb95-4dc8-8cb9-c7d70a806679') 'the quick brown fox']
[UUID('339619ab-bd17-401f-82c8-a927145d52cf')' jumps over the lazy dog']
]

Created a new array to store just the description values:

description = np.array([description[1] for description in chunk])

However, this gives me the following output:

['the quick brown fox','jumps over the lazy dog']

My desired output is:

[['the quick brown fox'],[jumps over the lazy dog']]

How can I achieve my desired output keeping the original order?

Using Python 3.8
The original list is from np.array_split(book_information, chunk_size) which I loop over in chunks

CodePudding user response：

You can simply add the additional dimension when defining your array: description = np.array([[description[1]] for description in chunk]).