approxEquals with relative tolerance #214
Comments
|
I think this would be a great addition. I’d just add it as another param rather than whole new methods personally. |
|
Ok. I think I can add that parameter without breaking source compatibility, but it will probably break binary compatibility. |
|
Not really, since it’s testing only there isn’t any real reason for someone
to need binary compatibility.
On Thu, Nov 9, 2017 at 12:53 AM Martin Mauch ***@***.***> wrote:
Ok. I think I can add that parameter without breaking source
compatibility, but it will probably break binary compatibility.
Is that something to worry about?
—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub
<#214 (comment)>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/AADp9ZYiGW-PKY11zu5ahYiDAm3TxVkUks5s0r1xgaJpZM4QWwR2>
.
--
Cell : 425-233-8271
|
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
In many cases, relative tolerances for approximate comparisons are more useful than absolute ones.
If e.g. my DataFrame contains some numbers in the range of
1.0E10and some in1.0E-10I would have to set a high absolute tolerance which would miss differences in the small numbers.If that is OK, I would create a PR that either adds a second
relativeTolparameter with default value, which gets only used iftol == 0.0or create a different set of methods altogether.The text was updated successfully, but these errors were encountered: