Inconsistent behavior of hexbins mincnt parameter, depending on C parameter #12926

Activity

1. dstansby commented on Apr 18, 2019

   

   dstansby

   on Apr 18, 2019

   Member

   More actions

   Sorry for the slow reply; if you can submit a PR that would be great!
2. github-actions commented on May 24, 2023

   

   github-actionsbot

   on May 24, 2023

   More actions

   This issue has been marked "inactive" because it has been 365 days since the last comment. If this issue is still present in recent Matplotlib releases, or the feature request is still wanted, please leave a comment and this label will be removed. If there are no updates in another 30 days, this issue will be automatically closed, but you are free to re-open or create a new issue if needed. We value issue reports, and this procedure is meant to help us resurface and prioritize issues that have not been addressed yet, not make them disappear. Thanks for your help!
3. github-actions

   added

   status: inactiveMarked by the “Stale” Github ActionMarked by the “Stale” Github Action

   on May 24, 2023
4. QuLogic commented on May 25, 2023

   

   QuLogic

   on May 25, 2023

   Member

   More actions

   Perhaps #18875 would fix these issues?
5. github-actions

   removed

   status: inactiveMarked by the “Stale” Github ActionMarked by the “Stale” Github Action

   on May 26, 2023
6. krooijers commented on Jun 12, 2023

   

   krooijers

   on Jun 12, 2023

   Last edited by krooijers

   ContributorAuthor

   More actions

   I've ran into this issue a couple more times over the past years.
   The inactivity notification was a good reminder to come back to this ...

   The issue is still present.
   With matplotlib 3.7.1 (and python 3.11.4 and numpy 1.24.3):
   

   I've slightly condensed the figure from my initial post, the code to produce the figure above is here

   from matplotlib import pyplot as plt
   import numpy as np
   
   plt.rcParams["font.size"] = 8
   plt.rcParams["font.family"] = "Noto Sans"
   
   np.random.seed(42)
   
   X, Y = np.random.multivariate_normal([0.0, 0.0], [[1.0, 0.1], [0.1, 1.0]], size=250).T
   Z = np.ones_like(X)
   
   extent = [-3., 3., -3., 3.]  # doc: "Order of scalars is (left, right, bottom, top)"
   gridsize = (7, 7)  # doc: "int or (int, int), optional, default is 100"
   hexbin_kwargs = dict(
       extent=extent,
       gridsize=gridsize,
       linewidth=0.0,
       edgecolor='none',
       cmap='Blues',
   )
   
   N_AXS = 6
   fig, axs = plt.subplots(int(np.ceil(N_AXS / 2)), 2, figsize=(8, 12))
   axiter = iter(axs.ravel())
   
   # call hexbin with varying parameters (C argument, mincnt, etc.):
   
   ax = next(axiter)
   ax.set_title("no mincnt specified, no C argument")
   ax.hexbin(
       X, Y,
       **hexbin_kwargs,
   )
   ax.set_facecolor("green")  # for contrast
   # shows a plot where all gridpoints are shown, even when the values are zero
   
   ax = next(axiter)
   ax.set_title("mincnt=1 specified, no C argument")
   ax.hexbin(
       X, Y,
       mincnt=1,
       **hexbin_kwargs,
   )
   ax.set_facecolor("green")
   # *all makes sense, so far*
   # shows only a plot where gridpoints containing at least one datum are shown
   
   ax = next(axiter)
   ax.set_title("no mincnt specified, C argument specified")
   ax.hexbin(
       X, Y,
       C=Z,
       reduce_C_function=np.sum,
       **hexbin_kwargs,
   )
   ax.set_facecolor("green")
   # shows only a plot where gridpoints containing at least one datum are shown
   
   ax = next(axiter)
   ax.set_title("mincnt=1 specified, C argument specified")
   ax.hexbin(
       X, Y,
       C=Z,
       reduce_C_function=np.sum,
       mincnt=1,
       **hexbin_kwargs,
   )
   ax.set_facecolor("green")
   # hmm, unexpected...
   # shows only a plot where gridpoints containing at least **two** data points are shown(!!!)
   
   ax = next(axiter)
   ax.set_title("mincnt=0 specified, C argument specified")
   ax.hexbin(
       X, Y,
       C=Z,
       reduce_C_function=np.sum,
       mincnt=0,
       **hexbin_kwargs,
   )
   ax.set_facecolor("green")
   
   # Highlight cells where sum == 0
   ax = next(axiter)
   ax.set_title("Cells where sum is zero (shaded black)")
   ax.hexbin(
       X, Y,
       C=Z,
       reduce_C_function=lambda v: sum(v) == 0,
       mincnt=-np.inf,
       **(hexbin_kwargs | dict(cmap="binary")),
   )
   ax.set_facecolor("green")

   @QuLogic #18875 does not improve things.. after installing https://github.com/MihaiBabiac/matplotlib/tree/bugfix/hexbin-marginals the figure renders the same. (and plt.matplotlib.__version__ == '3.3.2+1554.g54bf12686', FWIW)

   AFAICT the logic in _axes.py still needs to be revisited, although the code path is now slightly different compared to 2018.
7. oscargus

   closed this as completedin #26113on Jun 13, 2023
8. oscargus

   added a commit that references this issue on Jun 13, 2023

   Merge pull request #26113 from krooijers/bugfix-for-issue-12926

   ...

   This issue will close once commit 584cda0 is merged into the 'main' branch.

   584cda0
9. QuLogic

   added this to the v3.8.0 milestone on Jun 13, 2023
10. melissawm

    added a commit that references this issue on Jun 15, 2023

    Fixes matplotlib#12926 - inconsistency upon passing C in hexbin

    This issue will close once commit 720d692 is merged into the 'main' branch.

    720d692