@thosmos the :agency/id goes in the entity, which is a value in the env (an atom so it can be modified in a sideway and concurrently), the first one you send its an open link, can be one or many, but idents are always going to be a single entity, but I think you are complecting the data and the AST

the ident-reader moves the thing from the ident joins to context data, while the :agencies are your own resolver

but you should get the data using resolver inputs, are you having any trouble to deal with them this way?

I’m not having any trouble. I got it working now by using the env’s ::p/parent-query key as you suggested. But I realize I have a preference. My familiarity is from parsing EQL on the server side via Om/Next where the AST had everything I needed to process the request because it had info about the root key and the query keys, which included the ident key as the parent or root key, and the query keys as children. Currently, there’s insufficient information in the AST to process the request, since the AST only provides the first key from the query. I like that the ident is in the input of the resolver, but not that it’s missing from the AST. But, I don’t want to take any more or your time. I’ll just look at the code; if I want it to behave differently, I’ll figure out how to make it.

yeah, Pathom really tries to abstract a lot of that away from you, manipulations with AST should be used only for extensions and special cases, for regular things you shouldn't need it at all, that's the intended interface

ok got it