Overpass turbo: relations _inside_ area?

Question

This code

rel[name="Rheinisch-Bergischer Kreis"][admin_level=6][boundary=administrative];
out geom;

outputs my target area "Rheinisch-Bergischer Kreis" in germany. With this code

area[name="Rheinisch-Bergischer Kreis"][admin_level=6][boundary=administrative];
rel(area)[boundary="postal_code"];
out geom;

I'm trying to get those postal_code relations which are inside that target area. If you compare the two outputs, you can see that the selected area in the second query is bigger than the one in the first query. But it should be the same size.

There are little fractions of a relation on the west end that reach into my target area and therefore they're included in the output. How can I get only relations whose center is inside my target area? Or only relations which are fully part of my target area?

EDIT

Thank you mmd. I accept the dead end and now I present my "solution" (and my actual problem!), maybe this will help someone in the future:

I'm using this code

[timeout:900];

area[name="Nordrhein-Westfalen"][admin_level=4][boundary=administrative];

rel(area)[admin_level=5][boundary=administrative]; // regierungsbezirke

foreach(
  out geom;
  map_to_area;
  rel(area)[admin_level=6][boundary=administrative]; // kreise
  foreach(
    out geom;
    map_to_area;
    rel(area)[boundary="postal_code"];
    out geom;
  );
);

to get an OSM file, that outputs the admin_level 5 first, then admin_level 6, and then each postal code that belongs to that admin_level 6. My goal was to get a dictionary that maps postal code to admin_level 6 to admin_level 5. From the order inside the OSM file I can figure that out. Unfortunately the code above reports 37 postal codes more than once in the OSM file (grep for the tag -> print the postal_code (with awk e.g.) -> sort -> uniq -c -> sort -k 1,1). That's how I found out about the behaviour that I've described in my minimal-example-question. So what I'm gonna do now is to filter those 37 postal codes which appear multiple times inside the OSM file by hand. Then I can build my dictionary from the OSM file to get that mapping.

Answer 1

Regarding your initial issue, I'd suggest an approach which leverages existing nodes in the respective area. It assumes that every postal_boundary has at least one place node. If that's not the case for whatever reason, you could maybe use [amenity] instead. That's pretty much as close as it gets.

rel[name="Rheinisch-Bergischer Kreis"][admin_level=6][boundary=administrative];
map_to_area;
node(area)[place];
is_in;
rel(pivot)[boundary="postal_code"];
out geom;