What is the general rule of thumb when deciding between creating multiple schemas versus creating a single consolidated schema for different document types.
For example, if I want to index a collection of products and a collection of articles, what general rules should be followed to determine whether they should be created in one schema (and then use solr fq filter query to filter on the document type) or created in two schemas. The number of overlapping fields? The need to return data across both document types and also be able to filter to a single type?
There may not be any rule of thumb and is more of preference.
If you have entities
- which you want to show in your response together or
- have relationship between them
it would be better to have them as a Single index.
You can have different entities and want join them at query time it would help being in a single core. (Although with latest development, I think it is possible across cores as well)
If your entities are completely unrelated to each other it is better to have them as a separate core so that you maintain them differently.
Multiple cores can give you more flexibility to configure security at core level, variable incremental indexing and distribution for each core ....
Multiple cores may use some more resources depending upon the terms duplication, cache and so on