Task: Analyse the equity family against data-oriented design
Table of Contents
This page documents a task in the Rework the equity instruments to the instrument table rule story. It captures the goal, current status, acceptance, and any notes or results.
1. Goal
(Describe what user-visible-or-internal change this task produces.)
2. Status
| Field | Value |
|---|---|
| State | DONE |
| Parent story | Rework the equity instruments to the instrument table rule |
| Now | Nothing. |
| Waiting on | Nothing. |
| Next | Nothing. |
| Last touched | 2026-10-16 |
3. Acceptance
design.orgstates the family, its surfaces, the findings with the command behind every number, the proposed target model, the mapping from the current columns, and the staged units.design_er.pumldraws the proposed target model.- Every finding carries a severity, an effort and the data-oriented criterion it maps to.
- Each staged unit names the check that ends it.
4. Plan
(Implementation strategy. Written when work starts; key decisions
are distilled into the parent story's * Decisions at close, but the
plan itself stays — it is the historical record of what we did.)
5. Measurements
Every number this task claims about a tree, with the command that produced it
and the commit it was measured at. A number without both is not a measurement,
because a clean-up record describes a tree and trunk moves under a branch: a
consumer count, a gate's component count and a suite's size all change with no
code changing, and the reader of a bare number cannot tell which tree it came
from. Where the raw output matters, put it in an evidence/ file beside the
task and name that file in the Command column.
| What | Value | Command | Commit |
|---|---|---|---|
| Equity codes in the catalogue, and how many are routed | 32 total, 19 routed, 13 with no table | the awk in evidence/equity_catalogue_counts.txt |
d52d9b927e |
| Columns across the nine routed equity tables and their child | 218, from 17 to 26 per table | the per-table awk in the Measurements section of design.org | d52d9b927e |
6. Notes
7. Test Scenarios
Manual QA scenarios (scaffolded via compass add test_scenario)
that verify this task. Link
new ones here as they're created; the scenario doc itself links back
via its "Verifies task" field.
| Scenario | State | Notes |
|---|---|---|
8. PRs
9. Review
| # | Comment summary | File | Decision | Notes |
|---|---|---|---|---|
| 1 | Is variance_strike money or a continuous quantity? It is the only column whose classification the analysis left open. |
design.org, finding E3 | Accept as money | Ruled by the operator 2026-10-16. A variance strike is a level the parties agree and it settles a cash amount, so it is decimal and numeric(38,12). Recorded as a decision in design.org. |
| 2 | E5's "116 member names are reachable from one tagged write" is not what the number measures; it is the sum over all nine alternatives, and the M10 grep also counts each header's _economic_fields line. |
design.org, E5 and M10 | accept | Corrected: the nine alternatives declare 107 members; a tagged write reaches one alternative. M10 states how the 116 lines arise. |
| 3 | M15 names BOOST_AUTO_TEST_CASE, which returns 0; the 36 core cases are Catch2 TEST_CASE macros. |
design.org, M15 | accept | Command corrected. The 36 core, 172 shell and 29 ORE cases reproduce with TEST_CASE. |
| 4 | E14 says the positions have no top-level corpus document and appear only under a TotalReturnSwap. |
design.org, E14 | accept | Products/Input/referencedata.xml holds two top-level EquityPosition trades, and Hybrid_CFD.xml nests one under a ContractForDifference. E14 corrected. |
| 5 | M17's <SubTradeType>Equity= command returns 0, and its "2 nested" does not reproduce. |
design.org, M17 | accept | Corrected to 2 top-level, 5 files nesting EquityPosition and 3 nesting EquityOptionPosition. The corrected commands hold no pipe. |
| 6 | "32 columns are written by no mapper" ignores the generated synthetic-data generators under ores.trading.api/src/generators. |
design.org, E1, E7 and the use table | accept | A repo-wide search finds a generator writer for 20 of the 32 (all eight long_short, one settlement_type, rebate, payout_amount, price, four accumulator fields, the position currency, three Asian fields). The claim is scoped to the import mapper and the generators are named as a change surface. |
| 7 | The equity ORE round-trip gaps are owned by another story, so the design claims a gap it does not own. | design.org, the ORE-shape table and unit 10 | accept | Exotic_EquityAccumulator_single_name.xml, Exotic_EquityWorstOfBasketSwap.xml, Exotic_FXWorstOfBasketSwap.xml and FX_WorstOfBasketSwap.xml sit in known_export_gaps(). The missing range bounds and initial prices are owned by task 6605CCA2 in close-the-credit-loose-ends. Noted; unit 10 cannot pass until they close. |
| 8 | The hand-written import dispatcher and carrier builder are not named as surfaces the rework must touch. | design.org, "What each surface must change" | accept | ore_import_execute_handler.cpp copies every equity field into the NATS writes in a hand-written if constexpr chain, and instrument_batch_mapper.cpp builds the carrier. Both added. |
| 9 | The position fact table keeps currency, but the mapping drops it as writer-less. |
design.org, the fact tables and the mapping; design_er.puml | accept | No mapper sets equity_position_instruments.currency, and the design's own rule lands only mapper-written columns. The fact table now matches the mapping; the diagram still lists it and needs the same fix. |
| 10 | The child header calls weight a money-role value and the column an exact numeric. |
equity_position_option_underlying.hpp | decline | The design is right: a weight is continuous per the knowledge doc and the operator's ruling, and the column is double precision. The stale D11 comment is corrected in the code rework, not in the design. |
| 11 | The Result says the design records 33 findings. | the task page | accept | The findings table holds 16, E1 to E16. Result corrected. |
10. Result
The analysis of the equity family is written and the story is promoted from the product backlog into sprint 28.
design.org states the family, inventories its surfaces, records 16 findings
with the command behind each number, proposes the target model and stages the
rework into ten units. design_er.puml draws the proposal. The PNG is not
rendered yet; design.org names the command.
What the analysis found, in brief: the catalogue holds 32 codes of product type equity, 19 are routed, 17 of those to an equity table, and 13 have no instrument table at all. Nine routed models and one child table hold 222 columns, 17 to 26 per table, of which 17 of the 45 distinct business columns are assigned by no ORE import mapper. The family has no header. Regime 2 scores 1 of 5 and Regime 3 scores 1 of 5, both debt High.
The design was reviewed with the data-model review skill; its comments are in the Review table. It was re-measured after the rebase onto PR #3384, which added five equity columns, and the review corrected the finding count from 33 to 16 and declined three claims.
Merged as PR #3389.
Three follow-ups this task does not take.
- Ten staged units are named in the design and none is scaffolded as a task yet, so the story stays STARTED.
design_er.pngis not rendered; design.org names the command.- Running the build and the suites for the analysis turned up a defect
outside this family:
to_iso8601_utcformats throughto_time_tand"%Y-%m-%d %H:%M:%S", so it drops the whole fractional second, andparse_utc_timestampparses with the same seconds-only format and silently discards a fraction. A market observation therefore loses up to a second of its timestamp. The test that caught it isores.marketdata.service.tests. Taken as its own unit.