-
Notifications
You must be signed in to change notification settings - Fork 228
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
high centroid frequency causes overflow - giving incorrect results #93
Comments
Thanks Rohan!
…On Tue, Jul 25, 2017 at 11:52 AM, Rohan Sharma ***@***.***> wrote:
See sample code. When centroid frequency goes into billions and approaches
INT.MAX_VALUE, quantile method seems to go through an overflow and reports
incorrect result.
Data-structure state seems to be consistent - but when evaluating
quantile, incorrect results get returned.
[https://github.com/rohancs/t-digest-bug/blob/master/src/
main/java/TDigestBug.java]
—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
<#93>, or mute the thread
<https://github.com/notifications/unsubscribe-auth/AAPSet91royyb89llG5FWQSe3LWmd7Swks5sRjl2gaJpZM4Oi-ZH>
.
|
@rohancs Did you have a proposed fix? I am getting ready for a release. |
I ran your code on the master branch. I don't understand the problem or it has been fixed. I get this output from your code:
This seems OK to me. Can you comment? |
Hi @tdunning - sorry i was out on vacation and just got back. You are right - looks like this is fixed in the current master. My test was pinned against your last release version 3.1 |
That's great new.
I will resolve this issue and not block the next release.
…On Thu, Aug 3, 2017 at 12:52 PM, Rohan Sharma ***@***.***> wrote:
Hi @tdunning <https://github.com/tdunning> - sorry i was out on vacation
and just got back. You are right - looks like this is fixed in the current
master. My test was pinned against your last release version 3.1
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#93 (comment)>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/AAPSegr6B1PsB4aFsNxRNTqrCTFtUxodks5sUiT2gaJpZM4Oi-ZH>
.
|
sounds good. Thank you ! |
See sample code. When centroid frequency goes into billions and approaches INT.MAX_VALUE, quantile method seems to go through an overflow and reports incorrect result.
Data-structure state seems to be consistent - but when evaluating quantile, incorrect results get returned.
[https://github.com/rohancs/t-digest-bug/blob/master/src/main/java/TDigestBug.java]
The text was updated successfully, but these errors were encountered: