I don’t see what is surprising. The interface for a dictionary never specified the ordering of the keys, so I would not be surprised to see it vary based on release, platform, values of keys inserted, number of items in the dictionary, etc.
— Gerard Weatherby | Application Architect NMRbox | NAN | Department of Molecular Biology and Biophysics UConn Health 263 Farmington Avenue, Farmington, CT 06030-6406 uchc.edu On Aug 1, 2022, 7:48 PM -0400, Dan Stromberg <drsali...@gmail.com>, wrote: *** Attention: This is an external email. Use caution responding, opening attachments or clicking on links. *** On Mon, Aug 1, 2022 at 4:42 PM Dan Stromberg <drsali...@gmail.com> wrote: Yes, but I'm pretty sure that's been true for a LONG time. The hashes for small integers have been themselves for as long as I can remember. But the behaviour of the dictionary, when fed such keys, is what's changed. I'm not disputing either of those facts. I'm pointing out that the apparently arbitrary order of a mapping's keys becomes obvious when you look at the hashes of those keys. It looks like the relationship no longer holds at around keys = list(range(250, 260)) But i == hash(i) holds for the first million values at least. I could've been more clear. int dict keys stop being stored-in-order at near 256. But i == hash(i) holds for the first million values, and probably more. This suggests to me that there's something more than i == hash(i) going on inside dict's - but it doesn't much matter what it is for my purposes. -- https://urldefense.com/v3/__https://mail.python.org/mailman/listinfo/python-list__;!!Cn_UX_p3!j3p_Aq5MoGqDk5XMsKb4SKs3U1nfuMOx0wVkSa_hbURJ22w6lP8NrCOc_PYAfELYOdVlC9x6JzLfIMIw5sLe$ -- https://mail.python.org/mailman/listinfo/python-list
