ee.FeatureCollection to ee.Dictionary

[fitoprincipe]

#368 (review)

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 90.90%. Comparing base (a45bdb1) to head (5bc2ad0).
Report is 6 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #373      +/-   ##
==========================================
+ Coverage   90.85%   90.90%   +0.04%     
==========================================
  Files          27       27              
  Lines        1663     1671       +8     
  Branches       78       78              
==========================================
+ Hits         1511     1519       +8     
  Misses        130      130              
  Partials       22       22              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[12rambau]

As a general appreciation I love this PR as it solves an issue I open n the Google issue tracker a long time ago.

Appart from the comments I made on specific code, I'm asking myself a question, what error do we want to raise when the user specify a keycolumn that leads to duplicate id ?

1. not the same number of parameters ? then we go all the way to the Dictionary setup and and values and key will not have the same length
2. cannot use duplicate keys ? then same but we never use distinct on the keys

I think 2 is preferable (and that's what you coded) but I don't know if it was on purpose.

[fitoprincipe]

Answering to your general comment. Yes, it was on purpose. I actually wrote it down in the docstings. This is the perfect moment to express my follow up idea: I tried to applied a grouped reducer (ee.Reducer.group) in ee.FeatureCollection.reduceColumns but it didn't work. Although it's not complex to do it, I think a shortcut would be helpful, and could be an alternative way of doing what this PR is trying to implement. I will create a new PR with my idea.