Hi :)

We've been having some issues with huge product index that take a long time to build. That's because, we usually rely on the standard repositories supplied with Swift and Swift 2, and we could improve our house keeping routines on that and exclude fields or not use the ProductIndexSchemaExtender. The challenge, however, is to keep track of the usage of the fields, what queries and facets they are used in and also to keep the range of fields included/excluded up-to-date.

That's why I thought of a different approach. Today, we build repositories bottom-up, so we start with, what information should go i the index, then we build it an are able to work with the selection in queries. But what if the query setup was using the selected IndexSchemaExtender as the source for, which fields an be used in expressions? Then, the field selection in the index could be populated with, whatever fields were used in a query. Based on that, the builder would know, which index should be included, and which should be indexed. If we could also observe facets, the builder would know, which of the included fields should be stored and which should be analyzed.

That could reduce some of the issues we are experiencing (because of our bad habits), but it would also make the index management more wysiwyg'ish and make it more accessible to people with more frontend oriented skills since they would have a simpler workflow when setting up a facet / query.

Thanks for listening - I'll find my own way out😉

Br.
Lars