Pandas比较DataFrame行与前一行值时触发ValueError问题求助

尝试通过比较当前行与前一行的值创建新列时触发ValueError: The truth value of a DataFrame is ambiguous. Use a.empty, a.bool(), a.item(), a.any() or a.all()，已确认所有列数据类型为float64但错误仍存在。

相关代码 #

cols=['High', 'Low', 'Open', 'Volume', "Adj Close"]
df = df.drop(columns = cols)
df['EMA60'] = df['Close'].ewm(span=60, adjust=False).mean()
df['EMA100'] = df['Close'].ewm(span=100, adjust=False).mean()
df['MACD_60_100'] = df['EMA60'] - df['EMA100']
df['SIGNAL_60_100'] = df['MACD_60_100'].ewm(span=9, adjust=False).mean()
df['HIST_60_100'] = df['MACD_60_100'] - df['SIGNAL_60_100'] # Histogram
df = df.iloc[1: , :]  # Delete first row in DF as it contains NAN
print(df.dtypes)
print (df)
if df[df['HIST_60_100'] > df['HIST_60_100'].shift(+1)]: # check if the valus is > previous row value
    df['COLOR-60-100'] = "GREEN"
else:
    df['COLOR-60-100'] = "RED"

print(df.to_string())

报错信息 #

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
~\AppData\Local\Temp/ipykernel_10972/77395577.py in <module>
     32 
     33 
---> 34   get_data_from_yahoo(symbol+".NS")
     35 
     36   # df.to_excel(sheetXls, index=False)

~\AppData\Local\Temp/ipykernel_10972/520355426.py in get_data_from_yahoo(symbol, interval, start, end)
     26     print(df.dtypes)
     27     print (df)
---> 28     if df[df['HIST_60_100'] > df['HIST_60_100'].shift(+1)]:
     29         df['COLOR-60-100'] = "GREEN"
     30     else:

~\AppData\Roaming\Python\Python39\site-packages\pandas\core\generic.py in __nonzero__(self)
   1327 
   1328     def __nonzero__(self):
-> 1329         raise ValueError(
   1330             f"The truth value of a {type(self).__name__} is ambiguous. "
   1331             "Use a.empty, a.bool(), a.item(), a.any() or a.all()."

ValueError: The truth value of a DataFrame is ambiguous. Use a.empty, a.bool(), a.item(), a.any() or a.all().

df[df['HIST_60_100'] > df['HIST_60_100'].shift(+1)]返回的是过滤后的DataFrame对象，而非单个布尔值。Python的if语句只能判断单个布尔值，无法直接判断整个DataFrame的真假，因此触发歧义错误。

你的逻辑是想对每一行单独判断：如果当前行的HIST_60_100大于上一行，就标记为"GREEN"，否则"RED"，但当前写法是尝试把整个DataFrame放进if条件，不符合语法要求。

可以用以下两种方式实现需求：

方法1：使用numpy.where（推荐） #

通过向量化操作直接生成新列，效率更高：

import numpy as np

# 生成布尔序列：当前行 > 上一行
condition = df['HIST_60_100'] > df['HIST_60_100'].shift(1)
# 根据条件赋值，shift后第一行结果为NaN，会被视为False标记为RED
df['COLOR-60-100'] = np.where(condition, "GREEN", "RED")

方法2：使用apply逐行处理 #

如果需要更复杂的逻辑，也可以用apply遍历每一行：

import numpy as np

def determine_color(row, df):
    prev_val = df.loc[row.name - 1, 'HIST_60_100'] if row.name > 0 else np.nan
    return "GREEN" if row['HIST_60_100'] > prev_val else "RED"

df['COLOR-60-100'] = df.apply(determine_color, args=(df,), axis=1)

若需对第一行做特殊标记，可单独修改对应逻辑。

内容的提问来源于stack exchange，提问作者Ashish