Market-data lake

lake.backtest_proprietary_results dataset schema

Column types and research semantics for NexusTrade’s lake.backtest_proprietary_results dataset. Read the supported schema before submitting a bounded SQL query.

Query name
lake.backtest_proprietary_results
Grain
snapshot
Columns
27

Columns and types

These columns come from the same versioned logical catalog used by the query validator and engine. They describe the fields you can query. Some historical rows have missing values; fields absent from older shards can return typed NULL.

start_dateDATE
end_dateDATE
duration_daysINTEGER
date_range_bucketVARCHAR
intervalVARCHAR
cadenceVARCHAR
asset_classVARCHAR
tradedBOOLEAN
percent_changeDOUBLE
sharpe_ratioDOUBLE
sortino_ratioDOUBLE
max_drawdownDOUBLE
avg_drawdownDOUBLE
ulcer_indexDOUBLE
ulcer_performance_indexDOUBLE
win_rateDOUBLE
profit_factorDOUBLE
avg_trade_pnlDOUBLE
dollars_soldDOUBLE
total_dividendsDOUBLE
total_feesDOUBLE
risk_free_rateDOUBLE
baseline_percent_changeDOUBLE
baseline_sharpe_ratioDOUBLE
baseline_max_drawdownDOUBLE
alpha_pctDOUBLE
alpha_sharpeDOUBLE

Current logical schema for lake.backtest_proprietary_results

Coverage and timing

The logical view is backed by canonical manifest-resolved data. Query this name rather than guessing a physical database collection or object-storage path.

Backtest corpus: performance-only rows for runs on creator-owned data. NO IDENTIFIERS AND NO JOIN KEY, deliberately: no `backtest_uuid`, `portfolio_uuid` or `experiment_id`. It cannot be joined to any other table, and the strategies behind it cannot be identified or reconstructed. Never try to hydrate from it or link a row here to a row anywhere else. Without `experiment_id` the deduplication rule cannot be applied, so cloned runs count more than once. Treat any aggregate here as indicative and say so; never present it beside a deduplicated number as if the two were comparable.

Run a limited query

Use a registered API key with lake scope. Select needed columns, bind user-supplied values and restrict dates when the table has a time dimension. Query results are durable parts with a schema and manifest rather than an unbounded in-memory array.

This reference documents the schema and its meaning. Run queries in your signed-in workspace; this page does not execute SQL or show private results and datasets.

sql
SELECT "start_date", "end_date", "duration_days"
FROM lake.backtest_proprietary_results
LIMIT 20;