Fix: RTR sets INVESTIGATION.INV_STATE_CASE_ID fallback

[ericnagel]

Fix: Restore NULL behavior for missing investigation identifiers (STATE_CASE_ID bug)

Problem

PR #910 (June 22, 2026) inadvertently introduced a bug where contact investigations without real ODSE act_id rows were being assigned synthetic state case IDs derived from the investigation's local ID. This broke the expected MasterETL behavior where investigations without a real state case ID should have INV_STATE_CASE_ID = NULL.

Root Cause of PR #910 Bug

PR #910 added fallback logic to handle variable act_id_seq positions. While the Java transformer was improved to match identifiers by type_cd (not fixed positions), the SQL stored procedure injected a synthetic STATE act_id row:

UNION ALL
SELECT phc_fallback.public_health_case_uid AS [id],
       0 AS [act_id_seq],
       phc_fallback.record_status_cd AS [record_status],
       phc_fallback.local_id AS [root_extension_txt],  -- ← BUG: uses local_id as state case ID
       'STATE' AS [type_cd],
       'Local Public Health Case Identifier' AS [type_desc_txt],
       ...
FROM nbs_odse.dbo.public_health_case phc_fallback
WHERE phc_fallback.public_health_case_uid = phc.public_health_case_uid
  AND phc_fallback.local_id IS NOT NULL
  AND NOT EXISTS (SELECT 1 FROM nbs_odse.dbo.act_id ... WHERE ... type_cd = 'STATE')

This fallback was intended to handle edge cases but created fake identifiers for investigations like contacts that legitimately have no state case ID.

Solution

This PR removes the synthetic fallback behavior and restores the expected NULL handling through two complementary changes:

1. SQL Layer Fix (Stored Procedure)

Removed:

• The synthetic STATE fallback UNION ALL block
• The COALESCE that filled NULL root_extension_txt with phc.local_id for STATE rows

Result: Only real ODSE act_id rows are included in the JSON payload. Investigations without a real state case ID now correctly get NULL instead of a synthetic value.

2. Java Layer Defensive Filter

Added a safety check to ignore any synthetic or non-meaningful rows (those with act_id_seq <= 0):

// Ignore synthetic or non-meaningful rows such as fallback rows with act_id_seq = 0.
// Only real act_id rows should be considered for downstream case-id fields.
if (actIdSeq <= 0) {
  continue;
}

Testing

Added 8 new comprehensive unit tests covering missing scenarios, bringing total act_id transformation test coverage to 12 tests. Tests validate behavior for all combinations of identifier presence/absence and edge cases.

Test Coverage Matrix

Test Case	STATE	CITY	LEGACY	Description
testTransformActIdsExtractsByTypeCdRegardlessOfSequence	✅ seq 7	✅ seq 9	✅ seq 11	All present, non-standard sequences
testTransformActIdsIgnoresSyntheticFallbackStateRows	❌ seq 0	✅ seq 2	❌ absent	Synthetic rows (seq ≤ 0) filtered
testTransformActIdsLegacyCaseIdNullWhenNoLegacyRow	✅ seq 1	✅ seq 2	❌ absent	LEGACY missing → NULL
testTransformActIdsUsesLatestSequenceWhenMultipleLegacyRowsExist	❌ absent	❌ absent	✅ seq 30 (max)	Multiple rows → uses highest seq
testTransformActIdsEmptyArrayAllFieldsNull	❌ absent	❌ absent	❌ absent	CRITICAL: Contact investigations (0 rows)
testTransformActIdsStateNullWhenNoStateRow	❌ absent	✅ seq 1	✅ seq 2	STATE missing → NULL
testTransformActIdsCityNullWhenNoCityRow	✅ seq 1	❌ absent	✅ seq 3	CITY missing → NULL
testTransformActIdsMultipleStateRowsUsesHighestSequence	✅ seq 15 (max)	❌ absent	❌ absent	Multiple STATE rows → uses highest
testTransformActIdsMultipleCityRowsUsesHighestSequence	❌ absent	✅ seq 12 (max)	❌ absent	Multiple CITY rows → uses highest
testTransformActIdsBlankRootExtensionTxtIgnored	❌ blank	✅ "GA-CITY-001"	❌ blank	Whitespace/blank values skipped
testTransformActIdsAllTypesAbsentAllFieldsNull	❌ other	❌ other	❌ other	Unknown types only → all NULL
testTransformActIdsMixedPresenceAndAbsence	✅ seq 1	✅ seq 2	❌ absent	Mixed scenario validation

Bold rows = newly added tests

Impact

• ✅ Contact investigations now correctly have NULL identifiers instead of synthetic local_id values
• ✅ Restores MasterETL behavior and backward compatibility
• ✅ Fully tested with 100% coverage for all identifier type combinations
• ✅ Defensive filtering prevents any future synthetic row regressions

Related Issues

• APP-803
• Fixes: Investigations with no real ODSE act_id rows incorrectly assigned synthetic STATE_CASE_ID
• APP-717
• Related to: PR Investigation Identifier Fix — SQL Fallbacks & Fixture Normalization #910 (Investigation Identifier Fix)

Additional Notes

N/A

Checklist

• I have ensured that the pull request is of a manageable size, allowing it to be reviewed within a single session.
• I have reviewed my changes to ensure they are clear, concise, and well-documented.
• I have updated the documentation, if applicable.
• I have added or updated test cases to cover my changes, if applicable.

[ericbuckley]

any reason to not merge this fix into rel-7.13?

[ericnagel]

any reason to not merge this fix into rel-7.13?

@ericbuckley not that I can think of. I'll retarget it.