data_detail_problems and data_detail_video are empty in production #2868

Part of #2072.

Expected Behavior

data_detail_problems and data_detail_video have one row per problem or video event, like data_detail_nav does (about 30M rows). The Data Detail Problems and Data Detail Video charts on the Learner Engagement dashboard show data.

Current Behavior

Both tables in ol_warehouse_production_reporting have 0 rows (checked 2026-10-06), so both charts are blank.

Both models inner-join int__combined__course_runs on courserun_readable_id and platform (data_detail_problems.sql#L26-L30, data_detail_video.sql#L20-L24). tfact_problem_events and tfact_video_events use platform codes (mitxonline, edxorg, mitxpro, residential). int__combined__course_runs uses display names (MITx Online, edX.org, xPro, Residential MITx). The join never matches. data_detail_nav joins on courserun_readable_id only, so it isn't affected.

Steps to Reproduce

1. select count(*) from ol_warehouse_production_reporting.data_detail_problems returns 0.
2. select distinct platform from ol_warehouse_production_intermediate.int__combined__course_runs returns display names.
3. select distinct platform from ol_warehouse_production_dimensional.tfact_problem_events returns codes.

Possible Solution

int__combined__course_runs is only used here to look up course_title, and it is the only non-dimensional ref in either model. Replace it with dim_course_run (is_current, platform codes) joined to dim_course on course_fk, which fixes the join and finishes the #2072 migration for both models. #2686 did the same lookup swap in engagement_problem_completion_summary and handled the two SCD2 dims there.

• Remove both models' lines from dimensional_layering_baseline.txt.
• dim_course has no residential rows, so check course_title coverage per platform. dim_course_run.courserun_title is a possible fallback.