Ignore proj dicts with no key for slicing

[mraspaud]

• Tests added
• Tests passed
• Passes git diff origin/master **/*py | flake8 --diff

[coveralls]

Coverage increased (+0.1%) to 85.561% when pulling 2c20dc8 on fix-no-proj into b727fff on master.

[codecov]

Codecov Report

Merging #129 into master will increase coverage by 0.14%.
The diff coverage is 100%.

@@            Coverage Diff             @@
##           master     #129      +/-   ##
==========================================
+ Coverage   85.44%   85.59%   +0.14%     
==========================================
  Files          35       35              
  Lines        5957     5984      +27     
==========================================
+ Hits         5090     5122      +32     
+ Misses        867      862       -5

Impacted Files	Coverage Δ
pyresample/test/test_kd_tree.py	97.7% <100%> (+0.08%)	⬆️
pyresample/test/test_geometry.py	97.36% <100%> (+0.01%)	⬆️
pyresample/kd_tree.py	92% <100%> (+1.09%)	⬆️
pyresample/geometry.py	78.68% <100%> (ø)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update b727fff...2c20dc8. Read the comment docs.

[mraspaud]

this looks one level down I believe so it should be enough right ? ie in your case, if the variable being resampled has a coord called Time, we now look at the coords of Time itself.

[djhoese]

In the case of NUCAPS I believe I was getting errors because the coords had dimensions that the new DataArray did not. In NUCAPS this is something like Number_Of_FORs. So DataArray was complaining because I was making a DataArray with dims ('y', 'x') and coords that had dimensions that didn't exist. Let me see if I can reproduce it locally and I'll give you a self-contained example.

[djhoese]

ValueError: coordinate Time has dimensions ('Number_of_CrIS_FORs',), but these are not a subset of the DataArray dimensions ['y', 'x']

[mraspaud]

Isn't Number_of_CrIS_FORs a geographical dimension ?

[djhoese]

Technically for NUCAPS it is the only dimension. It is a time and location dimension. Everything is 1D in these NUCAPS EDR files. So Temperature, Time, Longitude, Latitude are all on the same dimension. Pressure is float Pressure(Number_of_CrIS_FORs, Number_of_P_Levels) ;.

[mraspaud]

well in that case it should be renamed to 'x' or 'y' so that pyresample understands it right ?

[djhoese]

Not without resampling the data because otherwise your coordinates would have different sizes than your output. We are resampling Temperature (Number_of_CrIS_FORs) to Temperature (y, x).

[djhoese]

Works for me with my NUCAPS dataset.