[SOLR-6966] For Data Driven Schema, consider multi-word text fields to be text, not string field types

• Type: Improvement
• Status: Closed
• Priority: Major
• Resolution: Duplicate
• Assignee: Unassigned
• Reporter: Grant Ingersoll
• Created: 12/Jan/15 16:24
• Updated: 08/Jun/17 21:20
• Resolved: 08/Jun/17 21:20

Description

A tricky situation, for sure, but I suspect in data-driven mode, when field guessing, we should treat multi-word strings as text by default, not String, so that the user's first experience is they can search against that field.

Alternatively, create a second field that is either the String version or the Text version.

Even more advanced option: use pseudo-fields (like what we do for some spatial) and intelligently use one or the other depending on the context: e.g. faceting uses the one form, search uses the other.

Issue Links

• duplicates: SOLR-9526 data_driven configs defaults to "strings" for unmapped fields, makes most fields containing "textual content" unsearchable, breaks tutorial examples
• is duplicated by: SOLR-7058 Data-driven schema needs to index large text fields as text and not as string
• is related to: SOLR-9526 data_driven configs defaults to "strings" for unmapped fields, makes most fields containing "textual content" unsearchable, breaks tutorial examples
• relates to: SOLR-8495 Schemaless mode cannot index large text fields