Interested in knowing more about this also.

@UPWHQK562 I’m having struggles with similar examples.

{::pco/output [:sub-item/id
                 {:sub-item/parent [:item/id]}]}

You may want to model the parent ID as a top-level attr of the sub-item record:

(pco/defresolver sub-item
  [{:sub-item/keys [id]}]
  {::pco/output [:sub-item/id :item/id]}
  (let [sub-items {1 {:sub-item/id 1 :item/id 1}
                   2 {:sub-item/id 2 :item/id 2}}]
    (get sub-items id)))

The other option is to model the :sub-item/parent attribute as a vector of parents rather than a singular parent record (note the [] around {:item/id 1} and {:item/id 2}:

(pco/defresolver sub-item
  [{:sub-item/keys [id]}]
  {::pco/output [:sub-item/id
                 {:sub-item/parent [:item/id]}]}
  (let [sub-items {1 {:sub-item/id 1 :sub-item/parent [{:item/id 1}]}
                   2 {:sub-item/id 2 :sub-item/parent [{:item/id 2}]}}]
    (get sub-items id)))

@UANMXF34G my understanding is that one of the limitations of EQL is that it doesn't specify whether the join is to a single value or a collection, and that {:x [:y]} is to-one or to-many depending on what is found. From Fulcro's docs: >Joins are automatically to-one if the data found in the state is a singular, and to-many if the data found is a vector. Is this a Fulcroism that doesn't apply to Pathom?

I don't know that much about Fulcro, so I'm not sure if this is just a Fulcro thing or a Pathom thing. It's surprising to me to hear that to-1 joins are possible with this syntax. However I see the EQL docs say the same thing! https://edn-query-language.org/eql/1.0.0/what-is-eql.html#_eql_for_selections Have you been able to traverse a to-1 join with Pathom before? I have never tried it, so it jumped out at me that you were trying it.

Afk right now so can't look at my code for exact examples, but I believe I have. 90% certain. Thank you for the suggestion but I believe it's inconclusive for now. I'll try to put some tests around this theory when I get in front of a computer. I would be surprised if it were the case, but it would be nice if a long held erroneous understanding got corrected.

I think I found the problem! Just a small typo:

(pco/defresolver item
 [{:item/keys [id]}]
 {::pco/output [[:item/id :item/foo]]}
 (get {1 {:item/id 1 :item/foo "a"}
       2 {:item/id 2 :item/foo "b"}}
      id))

(pco/defresolver item
 [{:item/keys [id]}]
 {::pco/output [:item/id :item/foo]}
 (get {1 {:item/id 1 :item/foo "a"}
       2 {:item/id 2 :item/foo "b"}}
      id))

The query works as written now:

(p.eql/process
  (pci/register [all-items
                 item
                 sub-item
                 earliest-sub-item])
  [{:all-items [:item/id
                ;; works!
                :item/foo
                {:item/earliest-sub-item [:sub-item/id
                                          {:sub-item/parent
                                           [:item/id :item/foo]}]}]}])
=>
{:all-items [#:item{:id 1, :foo "a", :earliest-sub-item #:sub-item{:id 1, :parent #:item{:id 1, :foo "a"}}}
             #:item{:id 2, :foo "b", :earliest-sub-item #:sub-item{:id 2, :parent #:item{:id 2, :foo "b"}}}]}

It was my own misunderstanding. Today I learned that we can express to-one queries in EQL!

Hah, oh man... Well, thank you. I wish there was same way to catch typos like that. I was looking for a typo in a misspelled key. OK, so now to continue with the repro of my actual issue lol I'll start another thread for that :)

I think the issue is that your response has only the value, you need each item on the collection to be like {:user/foo-times-two 5}

if its not, please provide a repro and we can check, mixing batch and non batch should be fine

@U066U8JQJ Thanks for the reply! I posted a repro in the GitHub of the issue that I am facing. Please let me know you need any more information. I apologize the code example above was incorrect (but in my repro, and in prod, I am indeed returning the full maps, not just the value) Link to the issue: https://github.com/wilkerlucio/pathom3/issues/171

replied at the issue