René's URL Explorer Experiment

{"@context":"https://schema.org","@type":"DiscussionForumPosting","headline":"poor categorical support w/ numpy\u003c1.8","articleBody":"Numpy versions \u003c 1.8 play poorly with arrays of mixed type, truncating anything not of type string to the length of the longest string. This is inhibiting #7974 \r\n\r\nSo the issue is as follows\r\n```python\r\n\u003e\u003e\u003e import numpy as np\r\n\u003e\u003e\u003e np.__version__\r\n'1.7.1'\r\n\u003e\u003e\u003e np.array(['A', 'A', np.nan, 'B', -np.inf, 3.14], dtype=np.unicode)\r\narray([u'A', u'A', u'n', u'B', u'-', u'3'], \r\n      dtype='\u003cU1')\r\n```\r\n\r\ntwo possible solutions:\r\n1) Fold back in shims removed in #7484 that existed mainly to solve this problem\r\n2) throw a warning if category.py is imported with a numpy\u003c1.8 and skip tests for numpy\u003c1.8\r\n\r\nThe advantage of 2 is it keeps the code cleaner and this is an edge case of sorts because it's only an issue when the array is of mixed type (strings and numbers). The advantage of 1 is of course edge case support on older systems. ","author":{"url":"https://github.com/story645","@type":"Person","name":"story645"},"datePublished":"2017-01-30T06:18:05.000Z","interactionStatistic":{"@type":"InteractionCounter","interactionType":"https://schema.org/CommentAction","userInteractionCount":3},"url":"https://github.com/7988/matplotlib/issues/7988"}