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/issues/7273#issuecomment-1411048944,https://api.github.com/repos/pydata/xarray/issues/7273,1411048944,IC_kwDOAMm_X85UGuXw,630436,2023-01-31T20:48:01Z,2023-01-31T20:48:01Z,NONE,"Sorry for slow reply, I got pulled onto another task before Christmas and forgot to revisit. Hit this in a different way again today. I get a combo of mangled as well as NaN values.


```
print((d0.swe.y - d1.swe.y).shape == d0.swe.shape)
print((d0.swe.y - d1.swe.y).shape)
print(d0.swe.y.shape)

False
(2509,)
(2509,)
```

```
np.unique(d0.swe.y.values == d1.swe.y.values)

array([ True])
```

```
np.unique(d0.swe.x.values == d1.swe.x.values)

array([False,  True]
```
```
xr.align(d0, d1, join=""exact"")
ValueError: cannot align objects with join='exact' where index/labels/sizes are not equal along these coordinates (dimensions): 'x' ('x',)
```

Looking at x shows X off by `1e-12`
```
np.unique(d0.swe.x.values - d1.swe.x.values)
array([-1.62003744e-12, -1.60582658e-12, -1.59161573e-12, -1.57740487e-12,
       -1.56319402e-12, -1.54898316e-12, -1.53477231e-12, -1.52056145e-12,
       -1.50635060e-12, -1.49213975e-12, -1.47792889e-12, -1.46371804e-12,
       -1.44950718e-12, -1.43529633e-12, -1.42108547e-12, -1.40687462e-12,
       -1.39266376e-12, -1.37845291e-12, -1.36424205e-12, -1.35003120e-12,
       -1.33582034e-12, -1.32160949e-12, -1.30739863e-12, -1.29318778e-12,
       -1.27897692e-12, -1.26476607e-12, -1.25055521e-12, -1.23634436e-12,
       -1.22213351e-12, -1.20792265e-12, -1.19371180e-12, -1.17950094e-12,
       -1.16529009e-12, -1.15107923e-12, -1.13686838e-12, -1.12265752e-12,
       -1.10844667e-12, -1.09423581e-12, -1.08002496e-12, -1.06581410e-12,
       -1.05160325e-12, -1.03739239e-12, -1.02318154e-12, -1.00897068e-12,
       -9.94759830e-13, -9.80548975e-13, -9.66338121e-13, -9.52127266e-13,
       -9.37916411e-13, -9.23705556e-13, -9.09494702e-13, -8.95283847e-13,
       -8.81072992e-13, -8.66862138e-13, -8.52651283e-13, -8.38440428e-13,
       -8.24229573e-13, -8.10018719e-13, -7.95807864e-13, -7.81597009e-13,
       -7.67386155e-13, -7.53175300e-13, -7.38964445e-13, -7.24753590e-13,
       -7.10542736e-13, -6.96331881e-13, -6.82121026e-13, -6.67910172e-13,
       -6.53699317e-13, -6.39488462e-13, -6.25277607e-13, -6.11066753e-13,
       -5.96855898e-13, -5.82645043e-13, -5.68434189e-13, -5.54223334e-13,
       -5.40012479e-13, -5.25801624e-13, -5.11590770e-13, -4.97379915e-13,
       -4.83169060e-13, -4.68958206e-13, -4.54747351e-13, -4.40536496e-13,
       -4.26325641e-13, -4.12114787e-13, -3.97903932e-13, -3.83693077e-13,
       -3.69482223e-13, -3.55271368e-13, -3.41060513e-13, -3.26849658e-13,
       -3.12638804e-13, -2.98427949e-13, -2.84217094e-13, -2.70006240e-13,
       -2.55795385e-13, -2.41584530e-13, -2.27373675e-13, -2.13162821e-13,
       -1.98951966e-13, -1.84741111e-13, -1.70530257e-13, -1.56319402e-13,
       -1.42108547e-13, -1.27897692e-13, -1.13686838e-13, -9.94759830e-14,
       -8.52651283e-14, -7.10542736e-14, -5.68434189e-14, -4.26325641e-14,
       -2.84217094e-14, -1.42108547e-14,  0.00000000e+00])
```
","{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,1441345452
https://github.com/pydata/xarray/issues/7273#issuecomment-1310901060,https://api.github.com/repos/pydata/xarray/issues/7273,1310901060,IC_kwDOAMm_X85OIsNE,10194086,2022-11-10T21:16:29Z,2022-11-10T21:16:29Z,MEMBER,"I suspect there are floating point inaccuracies which can sometimes trip up xarray. But I would also not expect differences when the coords are created with the same piece of code. 

Unfortunately your test does not work - it will first align and then do the subtraction... Any of the following should work:

```python
(d0.y - d1.y).shape == d0.shape
d0.y.values == d1.y.values
xr.align(d0, d1, join=""exact"")
```


If this worked earlier ","{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,1441345452
https://github.com/pydata/xarray/issues/7273#issuecomment-1309424575,https://api.github.com/repos/pydata/xarray/issues/7273,1309424575,IC_kwDOAMm_X85ODDu_,630436,2022-11-09T21:52:56Z,2022-11-09T21:52:56Z,NONE,"Each tiff is generated from the same underlying code so they are expected (and should be) exactly identical. I haven't seen any indication they are different. I just double checked and the two in the example above indicate they are the same
![image](https://user-images.githubusercontent.com/630436/200949548-9bc5dbf2-b7db-4e42-8bbd-251618c0ef04.png)

is xarray sensitive in a way that wouldn't be obvious with this kind of test?","{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,1441345452
https://github.com/pydata/xarray/issues/7273#issuecomment-1309335895,https://api.github.com/repos/pydata/xarray/issues/7273,1309335895,IC_kwDOAMm_X85OCuFX,10194086,2022-11-09T20:25:36Z,2022-11-09T20:25:36Z,MEMBER,Could it be that the coordinates of the individual arrays are not exactly the same?,"{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,1441345452
https://github.com/pydata/xarray/issues/7273#issuecomment-1308908697,https://api.github.com/repos/pydata/xarray/issues/7273,1308908697,IC_kwDOAMm_X85OBFyZ,630436,2022-11-09T15:10:30Z,2022-11-09T16:20:17Z,NONE,"Further investigation suggests this is similar / related to Issue #3681. Using `join=""override""` seems to resolve the problem. However, it is not at all clear why this needs to be done or why it ""fixes"" the problem. Is this intended behaviour?

edit: An old xarray 0.15.1 doesn't exhibit this behaviour and is correctly concated as expected. ","{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,1441345452