I am currently taking an online course on GIS and had a question about spatial joins. The course's example project is finding out which watersheds are in each county - there is a county feature class and a watersheds feature class. They describe their method as wanting to take attributes from counties to apply to watersheds so they have selected watersheds as the Target Feature and counties as the Join Feature.

I am working on a different spatial join exercise, this time involving finding precinct voting information per county. I am working with a precinct voting feature class and a county feature class.

My project only seems to have the ideal outcome when taking attributes from the precincts and applying it to counties, i.e. having counties as the Target Feature and precinct votes as the Join Feature.

I was wondering why that is, when the course example uses counties as the Join Feature?

1 Answer 1


First, precincts should never straddle county lines. Voting precinct boundaries often run through parks or the middle of streets where no one lives; you would not want to divide an apartment building into 2 precincts or, worse, two council districts. Districts are often created by grouping precincts.

Watersheds can and do cross counties. Counties can cross watersheds. So you could have a watershed in 2 or 3 counties; if you want to keep all the details of the watersheds, watersheds would be the one side (target) and counties, the join. Then you could pick up the county name and FIPS or other code from the intersecting county(ies). You can do this analysis in either direction. It is a question of which side you want to keep without any summarization. It is not unusual to look at this type of join/relate in both directions.

  • That's helpful! So a key in identifying which is the target is also whether these features overlap? In the precincts project, precincts should naturally be w/in counties so we would want to summarize precinct information & it wouldn't work vice-versa b/c counties are the more defined border? Whereas because watersheds & counties do overlap, we can do it both ways? It did seem that when I had precincts as the target feature instead of counties, a lot of information went missing or was calculated wrong.
    – GiGi
    Jan 26, 2022 at 19:27
  • It is not just overlap. It is the one-to-one, one-to-many, many-to-many relationships, which can happen without spatial relationships. The relationship of a precinct to a county (or a county to a state ) is almost a truth by definition. The precinct has to be smaller and it can't go outside its boundary. Some data like that, such as census data, often has attributes with codes indicating location. Other data takes much more analysis and exploration. The key is to explore your data and don't assume relationships.
    – danak
    Jan 27, 2022 at 0:05

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.