CLI reference

$ openqbw --help

Subcommand	What it does
catalog	Print the SYSTABLE catalog (table_id, name, root page).
schema	Print the columns of a table (via SYSCOLUMN bridged to SYSTABLE).
nulls	Histogram of SYSCOLUMN nulls_flag bytes with sample columns.
indexes	List SYSINDEX and cross-validate position attribution.
fkgraph	Print heuristic foreign-key edges (name-based fallback).
validate-attribution	Validate position attribution against SYSCOLUMN width bands.
export	Export transactions and line items to SQLite.
migrate	Export to CSV, SQLite, or IIF for data liberation.
forensics	File-level discovery report (pages, ap coverage, anomalies).
verify	Validate an export against known invariants.

Subcommand details

catalog

$ openqbw catalog mybooks.qbw

Prints one line per user table from SYSTABLE. Useful as a first look at any file.

schema <table>

$ openqbw schema mybooks.qbw abmc_invoice_lineitem

Lists the columns for the named table, bridging SYSCOLUMN to SYSTABLE via the table-id back-reference in the SYSTABLE row prefix (SPECIFICATION.md §5.1), and falling back to the SYSOBJECT catalog scan for owners that does not cover. The header line names the owner and the bridge that resolved it:

table: abmc_invoice_header  columns: 46  (SYSCOLUMN owner 3073, via SYSTABLE back-reference)

Not every table's columns are recoverable yet. When the recovered column_ids have holes the command says so on stderr, and when no owner can be bridged to the table it reports which stage came up empty - catalog presence, SYSCOLUMN recovery, or the bridge - rather than a bare "not found".

indexes [--fk-only] [--summary-only]

$ openqbw indexes mybooks.qbw
$ openqbw indexes mybooks.qbw --fk-only
$ openqbw indexes mybooks.qbw --summary-only

Lists every index from SYSINDEX (creator 0x01 0x46). With --fk-only, restricts to indexes whose name suggests a foreign key. With --summary-only, prints only the cross-validation result (agree / disagree / missing / orphan counts).

export --out <PATH>

$ openqbw export mybooks.qbw --out books.sqlite

Writes a SQLite database containing the catalog, lineitems, and transaction headers. The output is deterministic for the same input.

migrate --out <PATH> [--format csv|sqlite|iif]

$ openqbw migrate mybooks.qbw --out books.sqlite --format sqlite
$ openqbw migrate mybooks.qbw --out out_csv     --format csv
$ openqbw migrate mybooks.qbw --out books.iif   --format iif

Data-liberation export with three target formats:

• csv (default): writes catalog.csv, transactions.csv, and lineitems.csv into the directory given by --out (created if missing). Fields are RFC 4180 quoted only when needed.
• sqlite: alias for export --out <PATH>. Single deterministic SQLite database.
• iif: writes a single Intuit Interchange Format file with CRLF line endings. Line items are grouped by their parent invoice id; each group becomes one TRNS followed by SPL rows and an ENDTRNS. When a matching transaction header is available the header's transaction type is used, otherwise the group is emitted as GENERAL JOURNAL. SPL amounts are negated per IIF's double-entry convention.

forensics

$ openqbw forensics mybooks.qbw

File-level discovery report covering:

• File and page-store stats (size, page count, AP learned block coverage).
• Catalog summary (total tables, user tables).
• Business-record summary (transaction headers, line items, distinct parent ids, orphan parents, childless headers, lineitem grand total).

A non-zero orphan-parent count is a discovery signal: the file may contain partially purged records or a header table this build does not parse yet.

verify

$ openqbw verify mybooks.qbw

Runs the regression invariants:

• Invoice grand-total reconciliation (Phase 5)
• Per-table page-coverage check
• SYSINDEX cross-validation summary (WP-6Z.3)

Exit code is non-zero if any invariant fails.

validate-attribution

$ openqbw validate-attribution mybooks.qbw

Cross-checks the SYSTABLE-driven page->table map against the width-band attribution derived from SYSCOLUMN.

fkgraph

$ openqbw fkgraph mybooks.qbw

Falls back to a name-heuristic FK graph (*_id, *_id_h) for files where SYSINDEX is unparseable. Superseded by indexes when SYSINDEX is available.

nulls

$ openqbw nulls mybooks.qbw

Histogram of SYSCOLUMN.nulls_flag byte values with sample column names per value. Useful for understanding the null / default-marker discipline of a file.