Apply limit and offset after combining the relationships when the Content API gets multiple related/ values

Abierto
#38,594 0 comentarios 0 reacciones 0 asignados Ver en GitHub

Los mantenedores suelen responder en 2 días

Nadie ha tomado este issue todavía.

Evaluación

Dificultad
4/5
Tiempo estimado
3-5 días
Aptitud para principiantes
38/100
Tipo de issue
Error
Claridad
Bien especificado
Estado de actividad
Activo
Stack tecnológico
java
Área
api, search

Línea de trabajo

Start in ContentResource#getContent, where a related/ value with several relationships runs one query per relationship with the same limit and offset, then intersects the results. Reproduce with the steps in the issue, then check whether limit and offset can be applied once to the combined, sorted result. Done means the multi-relationship call returns old in the example, paging is correct, and an integration test covers the case.

Escrito por el modelo de indexación a partir del texto del issue.

Descripción

OKR : Customer Support Team : Maintenance Type : Defect

Problem Statement

A Content API call with more than one value in related/ returns too few results, or an empty list, when either relationship has more matching items than limit. The item matches every requested relationship and should be returned.

ContentResource#getContent handles related/A,B by running one query per relationship, each with the same limit and offset, and then keeping the items present in all of the lists. Each list is cut to the page size before the lists are intersected. An item that is not in the newest limit items of one relationship is dropped, even if it is the only item matching all the relationships. offset paging is broken the same way.

It fails silently: HTTP 200, an empty list, nothing in the log. Newer content added to a relationship can make an unchanged, published item disappear from the response. Re-saving or republishing the item moves it back up the modDate order, which looks like a fix but is temporary.

The Content API documentation ("Content API Retrieval and Querying") describes related/ with two fields as returning the content related to both, and limit as how many results are returned. The code comment on the intersection also says it filters the intersection.

Impact: every REST call that passes more than one related/ value is exposed. First seen on a self-hosted 25.07.10 LTS site.

Not an upgrade regression: the same logic is in 24.12.27 LTS and 25.07.10 LTS (and on main).

Suggested approach: apply limit and offset once, after combining the relationships. Either build a single query with all the relationship terms, or fetch each relationship unbounded, intersect, sort, then page.

Workaround: put the relationship terms in the Lucene query instead of related/, e.g. +Article.author:<id> +Article.topic:<id>.

Steps to Reproduce

https://drive.google.com/file/d/13E8TB1q8LHHcEQu-1qdRU2xttZd_joP1/view?usp=drive_link

  1. Create content types Author and Topic (a title each) and Article (a title plus two many-to-many relationship fields, author to Author and topic to Topic).
  2. Create and publish two Authors (A1, A2) and one Topic (T1).
  3. Create and publish one Article, old, related to A1 and T1.
  4. Create and publish 25 more Articles, newer than old, each related to A2 and T1.
  5. Call (anonymous):
    /api/content/render/false/query/+contentType:Article +live:true +languageId:1/limit/20/offset/0/orderBy/modDate desc/language/1/depth/0/related/Article.author:<A1>,Article.topic:<T1>
  6. Expected: 1 result (old). Actual: 0 results.
  7. For comparison (same data):
    • the same call with limit/100 returns old
    • only related/Article.author:<A1> with limit/20 returns old
    • only related/Article.topic:<T1> with limit/20 returns 20 items, not including old
    • no related/, with +Article.author:<A1> +Article.topic:<T1> in the query and limit/20, returns old
  8. Republish old, then repeat step 5: it returns old (its modDate is now the newest). It breaks again once 20 newer items share T1.

Acceptance Criteria

  • A call with several related/ values returns every item related to all of them, up to limit, regardless of how many items each relationship has
  • limit and offset apply once, to the combined and sorted result, so paging is correct
  • An integration test covers an item that matches two relationships while one relationship has more than limit newer items
  • The Content API documentation states how limit and offset behave with multiple related/ values

dotCMS Version

Reproduced on 25.07.10 LTS (v9) and 24.12.27 LTS (v2); same code on main.

Severity

Medium - Some functionality impacted

Links

Lenguaje dominante
Java
Estrellas
972
Forks
493
Merge medio
2 d 12 h
PR fusionados (30 d)
228

Preparar el entorno

Abrir en Codespaces

Inicia el contenedor de desarrollo del proyecto en tu navegador, con tu propia cuenta de GitHub.

Primeros pasos

  1. Lee el issue completo y luego la guía de contribución del proyecto.
  2. Comenta en el issue que vas a ocuparte — evita que dos personas hagan lo mismo.
  3. Haz un fork del repositorio y trabaja en una rama.
  4. Abre un pull request que haga referencia al número del issue.

Más de dotCMS/core

Todos los issues de dotCMS/core

Issues similares

Más issues de Java

Recibe los nuevos issues en tu correo

Un resumen breve de issues de GitHub para principiantes.