Datasets:

kangkangchen
/

a-share-qlib-context-600809

license: other
task_categories:
  - time-series-forecasting
tags:
  - finance
  - a-share
  - qlib
  - investment-data
pretty_name: A-share qlib context data for 600809.SH

A-share qlib Context Data for 600809.SH

This dataset is prepared from qlib-compatible data, especially the community data source chenditc/investment_data recommended by qlib while the official CN dataset is unavailable.

Source reference: https://github.com/chenditc/investment_data

Coverage

Stock: 600809.SH
Benchmark: 000300.SH
Date range exported: 20250102 to 20260424

Tables

data/qlib_daily: raw daily qlib fields available for the stock, such as open, high, low, close, volume, amount, factor.
data/qlib_factors_daily: derived daily context features computed from historical data only.
data/benchmark_daily: raw daily qlib fields for the benchmark, when provided.
data/outcome_targets: automatically computed future return and risk targets. These are outcome targets, not labels for main-force orders.

Price Adjustment and MA Check

For the chenditc/investment_data qlib provider, qlib open/high/low/close are modeling-oriented normalized feature prices. They should be kept with factor, and they should not be described as standard broker front-adjusted or back-adjusted prices without validation. This exporter keeps both concepts explicit:

feature_close = close
display_close = close / factor
ma5_display_close = rolling_mean(display_close, 5)

Sanity check from the local qlib data: for 山西汾酒 600809.SH on 2026-04-22, the five display closes ending that day are approximately 141.16, 139.50, 138.87, 136.86, 136.91, so ma5_display_close = 138.66. This matches the broker-style MA5 quoted by the user. Use qlib close/feature_close for modeling features, and use display_* fields when comparing to broker screens or business-facing price displays.

Leakage Policy

qlib_factors_daily only uses same-day and historical data. outcome_targets uses future prices and must only be used as supervised targets or evaluation fields.

There are no ground-truth labels for "main force orders" here. TWAP/VWAP/iceberg/protection patterns should be engineered from Level-2 data separately and joined by ts_code + trade_date.