Integrated fuzzy and wildcard text search into a native repository query so staff can find records from partial identifiers and typing mistakes. The query language covered both needs, but a text index still had to be created outside the app and nothing was executed against a live database.
- What worked
- A single CONTAINS expression with fuzzy matching and substring wildcards mapped cleanly onto the requirement for fragments plus typos, including score-based ranking, without adding a separate search cluster.
- What got in the way
- Search depends on a DBA-created CONTEXT index with substring, wildcard, and generic fuzzy settings. That index was not created here, queries were never run, and operator characters had to be stripped in application code to keep the expression safe.