Dear Solr Community,
I'm reaching out to ask for guidance regarding an issue I've encountered in Solr (version 8.11.4): I have a collection in which some documents previously included a field that now consistently has a NULL value after atomic update operation (I set <field_name>=NULL). However, I noticed that the field still appears to be present in the SegmentReader metadata, even though it's no longer actively used or populated and so it occupy space in RAM. My goal is to completely eliminate this field - not just from new documents, but also from the segment-level metadata - so that it no longer exists in the index at all (e.g., it should not show up in the schema field list or be referenced internally) in order to free heap memory. I cannot reindexing all documents into a new collection/core without the field and I experimented that merge or optimize aren't sufficient to purge it? I wanted to ask: - Is there a way to force Solr/Lucene to drop unused fields from SegmentReader without a full reindex? - Are there best practices for fully removing a field from an index if it was previously used? Any insights, tools, or strategies would be greatly appreciated. Thank you in advance for your time and support! Best regards, Antonio
