メニュー
Semantic model for the formula_1 database (Ergast F1 historical dataset — races, drivers, constructors, results, qualifying, pit stops, lap times). Load this skill when the user is querying the formula_1 database.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
Semantic model for the california_schools database. Contains business term definitions, metric formulas, known query patterns, and critical gotchas for the California Schools dataset (FRPM eligibility, SAT scores, district/school hierarchy). Load this skill when the user is querying the california_schools database.
Semantic model for the debit_card_specializing database (Czech fuel card transactions — gas stations, customers, products, monthly aggregates). Load this skill when the user is querying the debit_card_specializing database.
Semantic model for the european_football_2 database (Kaggle European Soccer dataset — matches, players, teams, leagues across 11 European countries). Load this skill when the user is querying the european_football_2 database.
Semantic model for the card_games database (Magic: The Gathering card data). Contains definitions for cards, sets, legalities, rulings, and foreign language data. Load this skill when the user is querying the card_games database.
Semantic model for the codebase_community database (StackOverflow-like Q&A community — posts, users, comments, votes, tags, badges). Load this skill when the user is querying the codebase_community database.
Semantic model for the financial database (Czech banking PKDD 1999 dataset). Contains business definitions for banking — loans, accounts, transactions, clients, districts. Load this skill when the user is querying the financial database.
SOC 職業分類に基づく
| name | formula-1-semantic-model |
| description | Semantic model for the formula_1 database (Ergast F1 historical dataset — races, drivers, constructors, results, qualifying, pit stops, lap times). Load this skill when the user is querying the formula_1 database. |
| metadata | {"author":"manual","version":"1.1","last_updated":"2026-02-25T00:00:00Z","learned_patterns":0,"verified_patterns":5} |
| Resource | What it contains | How to load |
|---|---|---|
| Extended gotchas | Time string vs milliseconds, NULL positions, positionText values, standings pitfalls | read_skill_resource("formula-1-semantic-model", "references/gotchas.md") |
| Metric definitions | Points totals, win counts, pit stop averages, championship standings | read_skill_resource("formula-1-semantic-model", "references/metrics.md") |
| Query patterns | Driver standings, constructor wins, fastest laps, DNF analysis | read_skill_resource("formula-1-semantic-model", "references/query_patterns.md") |
When to load resources:
references/gotchas.md (time columns are strings)references/gotchas.md (standings are cumulative)references/metrics.md| Term (user says) | Actual meaning | Confidence |
|---|---|---|
| "race winner" | results.position = 1 (integer — NULL for non-finishers) | verified |
| "did not finish" / "DNF" | results.position IS NULL or status.status not equal to 'Finished' | verified |
| "constructor" / "team" | constructors.name — joined via results.constructorId | verified |
| "grid position" | results.grid (starting position) — separate from finishing results.position | verified |
| "fastest lap" | results.rank = 1 (fastest lap rank within the race) | verified |
| "season" | races.year (integer, e.g. 2021) | verified |
| "race time" | results.milliseconds — use this for arithmetic; results.time is a formatted string | verified |
| "DNF reason" | status.status — joined via results.statusId → status.statusId | verified |
Time columns are strings — use milliseconds for arithmetic — results.time,
qualifying.q1/q2/q3, laptimes.time, and pitstops.duration are formatted
strings like "1:23.456". Only the milliseconds integer column is safe for
math and aggregation.
results.position is NULL for non-finishers — drivers who retired,
were disqualified, or did not start have NULL in position. Use positionOrder
(always populated) to rank all starters, or filter on status for DNF analysis.
driverstandings and constructorstandings are cumulative per race —
these tables store the running championship total after each race, not per-race
points. For the season total, filter to the last race of the season.
results.positionText contains non-numeric codes — values include 'R'
(retired), 'D' (disqualified), 'W' (withdrawn), 'F' (failed to qualify),
'N' (not classified). Never cast positionText to integer without guarding.
pitstops has both duration (text) and milliseconds (integer) —
always use milliseconds for pit stop duration math.
For extended gotchas and SQL examples, call:
read_skill_resource(
skill_name="formula-1-semantic-model",
resource_name="references/gotchas.md"
)
races.raceId ──→ results.raceId ──→ drivers.driverId
races.raceId ──→ results.raceId ──→ constructors.constructorId
races.raceId ──→ qualifying.raceId ──→ drivers.driverId
races.raceId ──→ laptimes.raceId ──→ drivers.driverId
races.raceId ──→ pitstops.raceId ──→ drivers.driverId
races.circuitId ──→ circuits.circuitId
results.statusId ──→ status.statusId (DNF reason)
races.raceId ──→ driverstandings.raceId
races.raceId ──→ constructorstandings.raceId
For canonical query patterns (season champions, win counts, fastest laps), call:
read_skill_resource(
skill_name="formula-1-semantic-model",
resource_name="references/query_patterns.md"
)
For full metric formulas (points per season, pit stop averages, constructor wins), call:
read_skill_resource(
skill_name="formula-1-semantic-model",
resource_name="references/metrics.md"
)