hsliuping/TradingAgents-CN · error · ValueError

DataFrame缺少收盘价列: {close_col}

Error message

DataFrame缺少收盘价列: {close_col}

What it means

add_all_indicators computes a full indicator suite (ma/rsi/macd/boll/atr/kdj columns) from a single close-price column, defaulting to close_col (typically 'close'). If that column is absent it raises immediately with this ValueError, since every downstream computation depends on it. Callers include the stock data tools (_format_stock_data, get_hk_stock_data_akshare), so raw data lacking a normalized 'close' column triggers it deep in formatting paths.

Source

Thrown at tradingagents/tools/analysis/indicators.py:318

        - ma5, ma10, ma20, ma60: 移动平均线
        - rsi: RSI指标（14日，国际标准）
        - rsi6, rsi12, rsi24: RSI指标（中国风格，仅当 rsi_style='china' 时）
        - rsi14: RSI指标（14日，简单移动平均，仅当 rsi_style='china' 时）
        - macd_dif, macd_dea, macd: MACD指标
        - boll_mid, boll_upper, boll_lower: 布林带

    示例：
        >>> df = pd.DataFrame({'close': [100, 101, 102, 103, 104]})
        >>> df = add_all_indicators(df)
        >>> print(df[['close', 'ma5', 'rsi']].tail())
        >>>
        >>> # 中国风格
        >>> df = add_all_indicators(df, rsi_style='china')
        >>> print(df[['close', 'rsi6', 'rsi12', 'rsi24']].tail())
    """
    # 检查必要的列
    if close_col not in df.columns:
        raise ValueError(f"DataFrame缺少收盘价列: {close_col}")

    # 计算移动平均线（MA5, MA10, MA20, MA60）
    df['ma5'] = ma(df[close_col], 5, min_periods=1)
    df['ma10'] = ma(df[close_col], 10, min_periods=1)
    df['ma20'] = ma(df[close_col], 20, min_periods=1)
    df['ma60'] = ma(df[close_col], 60, min_periods=1)

    # 计算RSI指标
    if rsi_style == 'china':
        # 中国风格：RSI6, RSI12, RSI24（使用中国式SMA）
        df['rsi6'] = rsi(df[close_col], 6, method='china')
        df['rsi12'] = rsi(df[close_col], 12, method='china')
        df['rsi24'] = rsi(df[close_col], 24, method='china')
        # 保留RSI14作为国际标准参考（使用简单移动平均）
        df['rsi14'] = rsi(df[close_col], 14, method='sma')
        # 为了兼容性，也添加 'rsi' 列（指向 rsi12）
        df['rsi'] = df['rsi12']
    else:

View on GitHub (pinned to 74783e8817)

Solutions

1. Rename the price column to match: df = df.rename(columns={'Close': 'close'}) or {'收盘': 'close'}.
2. Or pass the actual column explicitly: add_all_indicators(df, close_col='adj_close').
3. Ensure loaders (_format_stock_data / get_hk_stock_data_akshare paths) normalize column names before calling add_all_indicators.

Example fix

# before
df = add_all_indicators(raw_df)  # raw_df has '收盘'

# after
df = add_all_indicators(raw_df.rename(columns={"收盘": "close"}), close_col="close")

close_col = "close" if "close" in df.columns else next((c for c in ("Close", "收盘", "adj_close") if c in df.columns), None)
if close_col is None:
    raise ValueError("no price column found")
df = add_all_indicators(df, close_col=close_col)

def has_close_column(df: pd.DataFrame, close_col: str = "close") -> bool:
    """True if the DataFrame has the close column add_all_indicators needs."""
    return close_col in df.columns

try:
    df = add_all_indicators(df, close_col=close_col)
except ValueError as e:
    if "缺少收盘价列" in str(e):
        df = df.rename(columns={"Close": "close", "收盘": "close"})
        df = add_all_indicators(df, close_col="close")
    else:
        raise

When it happens

Trigger: Calling add_all_indicators(df) where df has 'Close', '收盘', 'adj_close', or no price column at all; passing close_col='adj_close' when only 'close' exists (or vice versa). Also hit indirectly via _format_stock_data or get_hk_stock_data_akshare on data whose columns were not normalized.

Common situations: Feeding DataFrames from different akshare endpoints with Chinese column names; renaming for storage and forgetting to map back; using adjusted vs raw close inconsistently across the pipeline.

Related errors

• DataFrame缺少必要列: {missing}, 现有列: {list(df.columns)[:10]}...
• 不支持的RSI计算方法: {method}，支持的方法: 'ema', 'sma', 'china'
• Indicator {indicator} is not supported. Please choose from:
• 不支持的指标: {name}

AI-assisted analysis of hsliuping/TradingAgents-CN@74783e8817 (2026-08-28). Data as JSON: /api/errors/5e7e7c70a79201b7. Report an issue: GitHub.