uproot5: Drawing of histograms created in uproot fails in ROOT

When saving a histogram in a ROOT file using Uproot, if the content of the last bin in the histogram is negative, ROOT will fail to plot it. The error message is

TCanvas::ResizePad:0: RuntimeWarning: Inf/NaN propagated to the pad. Check drawn objects.
TCanvas::ResizePad:0: RuntimeWarning: Canvas_1 height changed from 0 to 10

If the last bin’s content is non-negative, plotting works.

To reproduce, create the histogram

import numpy as np
import uproot

with uproot.recreate("test.root") as f:
    f["hist_fail"] = (np.asarray([-3, -2, -1]), np.asarray([0, 1, 2, 3]))
    f["hist_pass"] = (np.asarray([-3, -2, +1]), np.asarray([0, 1, 2, 3]))

Opening test.root using rootbrowser (or just root, getting the histograms, and plotting them using Draw()) results in the hist_pass being drawn, while hist_fail won’t be.

Furthermore, even the hist_pass isn’t drawn properly - the y-axis starts at 0. When creating the same histogram in ROOT, the range includes all the events.

I think the issue is caused by the errors of the histogram bins - ROOT initializes them as sqrt(abs(content)) while uproot uses sqrt(content), leading to nans for negative numbers. Not sure what is special about the last bin, though.

Uproot 4.3.4 ROOT 6.26/04

About this issue

Original URL
State: closed
Created 2 years ago
Comments: 17 (10 by maintainers)

Most upvoted comments

@jpivarski can we even write in the sumw2 information from uproot? because that’s needed, and right now it’s wrong

pywith(up.recreate("/tmp/example.root")) do file
        file["shit"] = np.array([-1,1,2]), np.array([0,1,2,3])
    end

#reading it back
fSumw2 => [0.0, -1.0, 1.0, 2.0, 0.0]

Moelf on Sep 1, 2022