René's URL Explorer Experiment

{"@context":"https://schema.org","@type":"DiscussionForumPosting","headline":"[Bug]: log scale on unitized data","articleBody":"### Bug summary\r\n\r\nNot totally sure if this is a bug, or a feature request, but for now should we maybe add a warning on trying to do log scale w/ dates \u0026 categories - i.e. non-float datatypes that the units system converts to floats but the resulting figure strips out all the unit information because it does the math on the underlying floats and doesn't convert back to the units. \r\n\r\n### Code for reproduction\r\n\r\n```python\r\nfig, ax = plt.subplots()\r\nax.plot(['a','b','c'], \r\n        [datetime(1999,12,1), datetime(2000,1,1), \r\n        datetime(2020, 10,1)])\r\nax.set_xscale('log')\r\nax.set_yscale('log')\r\n```\r\n\r\n\r\n### Actual outcome\r\n\r\n![unit](https://user-images.githubusercontent.com/1300499/131887631-6a4409cf-c72c-43da-b742-aa72e5db43c8.png)\r\n\r\n\r\n### Expected outcome\r\n\r\nprobably for the y-axis to still be labled in dates, and the x axis to be labeled in categories (which admittedly is impossible to do in this case) ","author":{"url":"https://github.com/story645","@type":"Person","name":"story645"},"datePublished":"2021-09-02T17:14:13.000Z","interactionStatistic":{"@type":"InteractionCounter","interactionType":"https://schema.org/CommentAction","userInteractionCount":24},"url":"https://github.com/20979/matplotlib/issues/20979"}