Data type of row object items

[jedwin3210]

Current course staff from Data 8 here,

We were wondering if returning "numpy.str_" was an intentional feature for type(tbl.row(...).item(...)). I bring this up since type(tbl.column(...).item(...)) returns a str. This means that in our test cases for jupyter notebooks, we should consider both of these types, but it would be beneficial to have both return a str for consistency and student readability.

[papajohn]

I don't think this behavior was intentional.

…

On Mon, Oct 28, 2024 at 5:54 PM Edwin Vargas Navarro < ***@***.***> wrote: Current course staff from Data 8 here, We were wondering if returning *"numpy.str_"* was an intentional feature for type(tbl.row(...).item(...)). I bring this up since type(tbl.column(...).item(...)) returns a str. This means that in our test cases for jupyter notebooks, we should consider both of these types, but it would be beneficial to have both return a str for consistency and student readability. Screenshot.2024-10-28.at.5.42.37.PM.png (view on web) <https://github.com/user-attachments/assets/4c0218fb-0d06-4471-b8f4-8ac42ee25be3> — Reply to this email directly, view it on GitHub <#618>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AACOFEKOZC25HGVJ4FFCILLZ53MF3AVCNFSM6AAAAABQYUFJROVHI2DSMVQWIX3LMV43ASLTON2WKOZSGYYTSOBWGU4TKNQ> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[davidwagner]

I'm open to a pull request to change/fix this.

[SebastianOuslis]

I think the core of this issue comes from the fact that the type returned by the row function and the column function are different.

This extends to the mismatch of types for other fields as well like integers.

I think the right thing to do would be to have both row and column return back the same type though this would be a potentially bigger scope of change than expected from the name of this issue.

[adnanhemani]

@SebastianOuslis - not sure that ensuring that both Rows and Columns having the same data types representing them is possible and/or efficient.

The main thing to think about here is: columns should have elements of the same data type - rows (in 99% of cases) will be varied data types.

But all in all, I agree - discussing that today (unless there is absolutely no better option) is likely not productive. Let's focus on any other quicker solutions that would solve this bug.

[liammaguire1]

Hi, first time contributor here. I'd like to check this out if it is still relevant.