Content inventory — TestBank_Backup_2026-11-21.json
Snapshot
mcqsubject values in the exportField coverage
| Field | Type | Populated (n of 2343) | Notes |
|-------|------|----------------------|-------|
| type | string | 2343 | All "mcq" |
| text | string | 2343 | Question stem |
| choices | array | 2343 | Array of {id, text} |
| correct | string | 2343 | Choice id |
| explanation | string | 485 | 1,858 empty (79%) |
| tables | array | 2 | Populated on only 2 questions |
| subject | string | 2343 | CPALE subject |
| topic | string | 2343 | Taxonomy topic |
| difficulty | string | 2343 | Enum: Easy, Moderate, Difficult, Very Difficult |
| tags | array | 2343 | May be empty array |
| id | string | 2343 | Format TB-000001 |
| dateAdded | string | 2343 | ISO timestamp |
| dateModified | string | 2343 | ISO timestamp |
| usage | object | 2343 | {timesUsed, lastUsed, testsUsedIn} |
| versions | array | 2343 | Recursive full snapshots |
| status | string | 2343 | Enum: draft, approved |
| archived | boolean | 2343 | 8 true |
| favorite | boolean | 2343 | User flag |
| journalEntries | array | 0 | Schema-only today |
| tAccounts | array | 0 | Schema-only today |
| schedules | array | 0 | Schema-only today |
| given | string | 0 | Schema-only today (empty string on all rows) |
| required | string | 0 | Schema-only today (empty string on all rows) |
| commonMistakes | string | 0 | Schema-only today (empty string on all rows) |
| standard | string | 0 | Schema-only today |
| standardParagraph | string | 0 | Schema-only today |
| bloomLevel | string | 2343 | Enum: "", Apply, Understand |
| collections | array | 2343 | May be empty array |
| chapter | string | 2197 | |
| subtopic | string | 2283 | |
| caseGroupId | string | 228 | References 57 distinct ids |
| nature | string | 2119 | Enum: "", None, Problem, Theory |
| course | string | 12 | |
| source | string | 12 | |
| reference | string | 12 | |
| author | string | 12 | |
| notes | string | 12 | |
| accountingArea | string | 2343 | Enum: "", Financial Accounting |
Eight fields with zero population (schema-only today): journalEntries, tAccounts, schedules, given, required, commonMistakes, standard, standardParagraph.
tables is populated on only 2 questions.
Enum domains
| Field | Observed values |
|-------|-----------------|
| difficulty | Easy, Moderate, Difficult, Very Difficult |
| nature | "", None, Problem, Theory |
| bloomLevel | "", Apply, Understand |
| status | draft, approved |
| accountingArea | "", Financial Accounting |
Note: empty string is a real value in these columns.
Data-integrity findings
1. 45 of 57 case group ids referenced by questions have no entry in caseGroups — the scenario text is missing, so those item sets cannot be rendered. Example dangling ids: CASE-0fwkrqv, CASE-0ijbmzn, CASE-0tfd4g2, CASE-1t2yw9s, CASE-1y1h0jx.
2. 1,858 of 2,343 questions (79%) have an empty explanation.
3. Eight accounting-presentation fields are populated on zero questions — the journal-entry / t-account / depreciation-schedule differentiators are unimplemented in content (journalEntries, tAccounts, schedules, given, required, commonMistakes, standard, standardParagraph).
4. 9 topics in the taxonomy are non-accounting demo leftovers — Chemistry, Geography, Geometry, History of Science, Orientation, Philippine Government, Philippine History, Physics, Workplace Ethics — with no questions attached.
5. Three of the six LECPA subjects have no content — Advanced Financial Accounting and Reporting, Auditing, and Regulatory Framework for Business Transactions. The export's four subject values are Financial Accounting and Reporting, MANAGEMENT SERVICES, TAXATION and Theory of Accounts; only the first three are LECPA subjects, and Theory of Accounts needs a mapping decision at import.
Consequences for the build
All three headline findings are resolved by decisions recorded in docs/ADR-004-content-decisions.md
(subject mapping, case-group grouping, explanation handling).
caseGroupId should be flagged or the missing scenarios must be supplied.How to reproduce
These figures come from reading the JSON with a JSON parser and counting keys; the orchestrator's script is the source.