issue_comments: 363528599
This data as json
| html_url | issue_url | id | node_id | user | created_at | updated_at | author_association | body | reactions | performed_via_github_app | issue |
|---|---|---|---|---|---|---|---|---|---|---|---|
| https://github.com/pydata/xarray/pull/1885#issuecomment-363528599 | https://api.github.com/repos/pydata/xarray/issues/1885 | 363528599 | MDEyOklzc3VlQ29tbWVudDM2MzUyODU5OQ== | 10050469 | 2018-02-06T19:00:45Z | 2018-02-06T19:01:21Z | MEMBER | Which, I guess, illustrates our point quite well: while it's easy for users to find out when their data is messy, it's much harder to find a robust and logical way to handle all cases for a general library like xarray. I guess it's the reason why matplotlib also doesn't warn or error when unordered data is given to pcolormesh. Don't get me wrong: I think it would be valuable to be able to provide a better "service" than matplotlib. Probably we should even error earlier, at the 2d plot level (and not only at the pcolormesh level like it is now). It's just going to be a bit hard because of all the special cases (dtypes, 2d coords, and map projections moslty). Ordering "per default" might be the most straightforward way, but: - what about performance issues? - I don't like the idea of silently doing things with user data I'll let @shoyer decide on what's best here. I am slightly in favor of doing nothing and giving the responsibility to the user. |
{
"total_count": 0,
"+1": 0,
"-1": 0,
"laugh": 0,
"hooray": 0,
"confused": 0,
"heart": 0,
"rocket": 0,
"eyes": 0
} |
294089233 |