The main idea is to create an index as a property of the list from the text (exclude and remove stop words). To improve performance, use “relationship indexing” by moving the list property to a child. This prevents the loading of a potentially large list as part of the default fetch group - you only need to request a query. You will need to use low-level api to execute only key requests. Taht will return the keys of the parent class, which can then be used to get the corresponding elements.