[naseweisssss]

#167 Giving credit to this PR, but added more test , this is intended to resolve #166

[seberg]

Thanks, LGTM... I was hesitant a bit wondering if we can avoid the result_type here, but it seems hard without anything else.
(And even if, it would seem to require quite a refactor.

I'd love if someone else would have a look, but that seems unlikely here :)

[naseweisssss]

Thanks, LGTM... I was hesitant a bit wondering if we can avoid the result_type here, but it seems hard without anything else. (And even if, it would seem to require quite a refactor.

I'd love if someone else would have a look, but that seems unlikely here :)

Thanks for approving but seems like it requires some other access to do this here.

[AbhiRKeesara]

@naseweisssss Thanks for this PR and mentioning of PR from kritagyadubey for credit.

A small ask if it's ok. Can we also add his name as co-contributor in the description if his changes have helped instead of just the PR reference?

@kritagyadubey - If you can, would you mind reviewing this PR before its merged?

[kritagyadubey]

@AbhiRKeesara thanks for the nudge - happy to review.

I went through the changes and I'm happy with them, so approving here. The fix is the same idea as what I had in #167 - taking the result dtype from all the inputs instead of just rate - and the tests here go further than mine, especially the ppmt/ipmt cases at 0% and the float32/Decimal checks. I double-checked the expected numbers in the new tests and they all come out right. CI is green as well.

One small optional suggestion: a short comment above the empty_like line saying why the dtype comes from all the inputs would help stop it from being simplified back later.

Thanks @naseweisssss for taking this on and for crediting #167 - good to merge from my side.

[kritagyadubey]

I went through the changes and I'm happy with them, so approving here. The fix is the same idea as what I had in #167 <#167> - taking the result dtype from all the inputs instead of just rate - and the tests here go further than mine, especially the ppmt/ipmt cases at 0% and the float32/Decimal checks. I double-checked the expected numbers in the new tests and they all come out right. CI is green as well. One small optional suggestion: a short comment above the empty_like line saying why the dtype comes from all the inputs would help stop it from being simplified back later. Thanks @naseweisssss <https://github.com/naseweisssss> for taking this on and for crediting #167 <#167> - good to merge from my side.

…

On Fri, Oct 23, 2026 at 3:33 AM Abhinay Reddy Keesara < ***@***.***> wrote: *AbhiRKeesara* left a comment (numpy/numpy-financial#170) <#170 (comment)> @naseweisssss <https://github.com/naseweisssss> Thanks for this PR and mentioning of PR from kritagyadubey <https://github.com/kritagyadubey> for credit. A small ask if it's ok. Can we also add his name as co-contributor in the description if his changes have helped instead of just the PR reference? @kritagyadubey <https://github.com/kritagyadubey> - If you can, would you mind reviewing this PR before its merged? — Reply to this email directly, view it on GitHub <#170?email_source=notifications&email_token=CHXX2HOXNECQCP7WDAWVEUD5VJ7ZTA5CNFSNUABFM5UWIORPF5TWS5BNNB2WEL2JONZXKZKDN5WW2ZLOOQXTMMZVGIZDQOBXGMZKM4TFMFZW63VHNVSW45DJN5XKKZLWMVXHJLDGN5XXIZLSL5RWY2LDNM#issuecomment-6352288732>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/CHXX2HL4PDAMM3WYNP5HZOD5VJ7ZTAVCNFSNUABFKJSXA33TNF2G64TZHMZDCMBZGI2TENJRHNEXG43VMU5TKOJTGEYDANBQGUZKC5QC> . You are receiving this because you were mentioned.Message ID: ***@***.***>