dakshana_fact_reported_results: no student id — name-only key loses ~29% of matches

**Problem.** `dakshana_fact_reported_results` keys students by **name only** — no `student_id` / `application_no` / DoB. (`drn` is present on the JEE-Main sheet only, blank for NEET, and doesn't join to Avanti records.)

**Why it matters.** Attaching these scholars to Avanti/golden records by name+cohort recovers only **~71%**. For 2025: JNV Bengaluru Urban 54/76 matched, JNV Pune 80/118 matched. The unmatched ~29% carry real qualified results that are lost to any per-centre outcome rollup — so Dakshana centre outcomes are systematically undercounted.

**Open question for the data team.** Can a stable identifier be added at ingestion — ideally the NTA application number or roll number (which would also let it join to `jnv_fact_jee_results` / `fact_student_*`), or failing that DoB alongside name to disambiguate?

Related: golden-build now surfaces Dakshana `jee_main_qual`/`neet_qual` from this table by name match, so it directly benefits from a better key.

Start by tracing ingestion for dakshana_fact_reported_results and compare its student fields with jnv_fact_jee_results and fact_student_*; check whether an NTA application or roll number, or DoB, is available before the name-only join. Done means a stable key or documented fallback is ingested, and the 2025 centre matches and golden-build qualification rollups recover the currently unmatched records.

Written by the indexing model from the issue text.