financial-analyst-data-full
A-share historical price + valuation data packaged for financial-analyst — the 14-agent single-stock deep-dive research workstation.
Published: 2026-05-24
Preset: full — 全 A 股完整包 (含历史退市股). 量化研究员 / 重度用户. lite 全 + TDX 历年财报原始 zip (用户跑 import_tdx_financial.py 解) + F10 原始文本 (公司大事/龙虎榜/主力追踪/最新提示 .txt).
Size: ~14.1 GB
What's included
- 5450 stocks daily OHLCV + 7 valuation fields (PE/PB/PS/DV/MV/CIRC_MV/turnover_rate)
- Date range (daily): 1990-12-19 → 2026-12-31 (8797 trading days)
- 5min OHLCV for the same stocks (rolling ~7-100 days)
- 24 parquet files (748.6 MB): industry / concept / index_constituents / TDX F10 / northbound / financials
- TDX 历年财务原始 zip (117 files, 256.6 MB) — 用
import_tdx_financial.py解 - F10 原始文本 (22090 files, 294.1 MB) — 公司大事/龙虎榜单/主力追踪/最新提示 .txt, agent 读详情用
- Two binary formats:
- Qlib
.bin(time-series,[4-byte float32 start_idx] + [float32 array]) for OHLCV + factors - Apache Parquet (columnar) for financials / events / institutional / industry mappings
- Qlib
- Compatible with Microsoft Qlib and financial-analyst loaders
Directory layout
cn_data/ # daily
calendars/day.txt
instruments/all.txt # 5450 codes
features/{code}/ # e.g. sh600519/
open.day.bin
high.day.bin
low.day.bin
close.day.bin
volume.day.bin
amount.day.bin
pe_ttm.day.bin
pb.day.bin
ps_ttm.day.bin (may be NaN if Tushare-free source)
dv_ttm.day.bin (may be NaN)
total_mv.day.bin (单位: 万元)
circ_mv.day.bin (单位: 万元)
turnover_rate.day.bin (单位: %)
cn_data_5min/ # 5-minute, similar layout
parquet/ # 非时序结构化数据 (列存)
industry_boards.parquet # 同花顺一级行业板块
index_constituents.parquet # CSI300/CSI500 等指数成份
tdx_f10_index.parquet # F10 事件索引 (公司大事/龙虎榜/研报)
tdx_f10_warnings_latest.parquet # 最新负向预警 (utility 警示)
northbound_holding.parquet # 北向资金持仓
tushare_stock_basic.parquet # 股票基本信息
concept_ths_*.parquet # 同花顺概念
events/ # 公司事件 (公告)
institutional/ # 机构持仓
blocks/ # 板块映射
xdxr/ # 分红除权
financial/ # 完整财务报表 (~735MB)
tdx_finance/ # TDX 历年财报原始 zip (用户自解)
Usage
Option 1 — via financial-analyst CLI (recommended)
pip install financial-analyst
fa init # interactive wizard, picks up this dataset
fa report SH600519
Option 2 — direct download + Qlib
from huggingface_hub import snapshot_download
local_dir = snapshot_download(
repo_id="jesson-hh/financial-analyst-data-full",
repo_type="dataset",
local_dir="~/.financial-analyst/data",
)
import qlib
qlib.init(provider_uri="~/.financial-analyst/data/cn_data", region="cn")
from qlib.data import D
df = D.features(["SH600519"], ["$close", "$volume"],
start_time="2024-01-01", end_time="2026-05-30", freq="day")
Units & conventions
| Field | Unit | Notes |
|---|---|---|
| open / high / low / close | 元 (CNY) | not adjusted (前复权) — adjustment factor not included |
| volume | 手 (= 100 股) | Tushare convention; NOT pytdx convention (股) |
| amount | 元 | 成交额 |
| pe_ttm / pb / ps_ttm | (无单位) | TTM ratios |
| dv_ttm | % | dividend yield TTM |
| total_mv / circ_mv | 万元 | total / circulating market cap |
| turnover_rate | % | daily turnover ratio |
Data sources & lineage
- OHLCV + financials: Tushare Pro (
pro.daily+daily_basic), HTTP endpoint - 5min OHLCV: TDX main sites via pytdx
- For daily updates (post-download): users run
fa data updatewhich pulls from pytdx main sites (free, no token) + Tencent qt.gtimg.cn realtime (free, no cookie). See docs/research/2026-05-23-direct-data-stability.md.
Disclaimer
This dataset is for research and educational purposes only. Data accuracy not guaranteed; do not use for trading without independent verification. See Tushare ToS for redistribution terms.
License
Apache 2.0 (code) — refer to original sources (Tushare / TDX) for data licensing.
Generated by scripts/publish_hf_dataset.py v1.