Pandas样式应用中Boolean Mask异常行为的问题排查

处理包含<x格式的数据时，希望将这类值转换为x/2（例如<5转为2.5），同时为修改过的值添加样式。数据列包含混合的数值与文本类型，使用Boolean Mask标记需要修改的单元格，转换函数运行正常，但样式应用时出现异常：HCO3列中原本为文本型0.5的未修改值、CO3列的非NaN值被错误地格式化。

测试数据与代码如下：

import pandas as pd
import numpy as np

dummy={'Location': {0: 'Perth', 1: 'Perth', 2: 'Perth', 3: 'Perth', 4: 'Perth', 5: 'Perth', 6: 'Perth', 7: 'Perth', 8: 'Perth', 9: 'Perth', 10: 'Perth', 11: 'Perth', 12: 'Perth', 13: 'Perth', 14: 'Perth', 15: 'Perth', 16: 'Perth', 17: 'Perth'}, 'Date': {0: '11/01/2012 0:00', 1: '11/01/2012 0:00', 2: '20/03/2012 0:00', 3: '6/06/2012 0:00', 4: '14/09/2012 0:00', 5: '17/12/2013 0:00', 6: '1/02/2014 0:00', 7: '1/02/2014 0:00', 8: '1/02/2014 0:00', 9: '1/02/2014 0:00', 10: '1/02/2014 0:00', 11: '1/02/2014 0:00', 12: '1/02/2014 0:00', 13: '1/02/2014 0:00', 14: '1/02/2014 0:00', 15: '1/02/2014 0:00', 16: '1/02/2014 0:00', 17: '1/02/2014 0:00'}, 'As µg/L': {0: '9630', 1: '9630', 2: '8580', 3: '4990', 4: '6100', 5: '282', 6: '21', 7: '<1', 8: '<1', 9: '<1', 10: '<1', 11: '<1', 12: '<1', 13: '<1', 14: '<1', 15: '<1', 16: '<1', 17: '<1'}, 'As': {0: '9.63', 1: '9.63', 2: '8.58', 3: '4.99', 4: '6.1', 5: '0.282', 6: '0.021', 7: '<1', 8: '<1', 9: '<1', 10: '<1', 11: '<1', 12: '<1', 13: '<1', 14: '<1', 15: '<1', 16: '<1', 17: '10'}, 'Ba': {0: 1000.0, 1: np.nan, 2: np.nan, 3: np.nan, 4: np.nan, 5: np.nan, 6: np.nan, 7: np.nan, 8: np.nan, 9: np.nan, 10: np.nan, 11: np.nan, 12: np.nan, 13: np.nan, 14: np.nan, 15: np.nan, 16: np.nan, 17: np.nan}, 'HCO3': {0: '10.00', 1: '0.50', 2: '0.50', 3: '<22', 4: '0.50', 5: '0.50', 6: '0.50', 7: np.nan, 8: np.nan, 9: np.nan, 10: '0.50', 11: np.nan, 12: np.nan, 13: np.nan, 14: np.nan, 15: np.nan, 16: np.nan, 17: np.nan}, 'Cd': {0: 0.0094, 1: 0.0094, 2: 0.011, 3: 0.0035, 4: 0.004, 5: 0.002, 6: 0.0019, 7: np.nan, 8: np.nan, 9: np.nan, 10: np.nan, 11: np.nan, 12: np.nan, 13: np.nan, 14: np.nan, 15: np.nan, 16: np.nan, 17: np.nan}, 'Ca': {0: 248.0, 1: 248.0, 2: 232.0, 3: 108.0, 4: 150.0, 5: 396.0, 6: 472.0, 7: np.nan, 8: np.nan, 9: np.nan, 10: np.nan, 11: np.nan, 12: np.nan, 13: np.nan, 14: 472.0, 15: np.nan, 16: np.nan, 17: np.nan}, 'CO3': {0: 0.5, 1: 0.5, 2: 0.5, 3: 0.5, 4: 0.5, 5: 0.5, 6: 0.5, 7: np.nan, 8: np.nan, 9: 0.5, 10: np.nan, 11: np.nan, 12: np.nan, 13: np.nan, 14: np.nan, 15: np.nan, 16: np.nan, 17: np.nan}, 'Cl': {0: 2.0, 1: 2.0, 2: 2.0, 3: 2.0, 4: 0.5, 5: 2.0, 6: 5.0, 7: np.nan, 8: np.nan, 9: np.nan, 10: np.nan, 11: np.nan, 12: np.nan, 13: 5.0, 14: np.nan, 15: np.nan, 16: np.nan, 17: np.nan}}

df=pd.DataFrame(dummy)

mask = df.applymap(lambda x: (isinstance(x, str) and x.startswith('<')))
def remove_less_thans(x):
    if type(x) is int:
        return x
    elif type(x) is float:
        return x
    elif type(x) is str and x[0]=="<":
        try:
            return float(x[1:])/2
        except:
            return x
    elif type(x) is str and len(x)<10:
        try:
            return float(x)
        except:
            return x
    else:
        return x
def colour_mask(val):
    colour='color: red; font-weight: bold' if val in df.values[mask] else ''
    return colour

# 执行转换逻辑
df=df.applymap(remove_less_thans)

styled_df= df.style.applymap(colour_mask)
styled_df

1. mask时效性失效：mask基于原始数据生成，但后续对df执行了转换操作修改了数据内容，此时df.values[mask]指向的是修改后的数据中原始mask标记位置的值，而非原始的<x字符串。
2. 值匹配逻辑错误：val in df.values[mask]是全局值匹配，而非基于单元格位置判断。例如原始<1转换后为0.5，而HCO3列的0.50转换后也是0.5、CO3列原本就存在0.5，这些重复值会被错误判定为需要高亮的对象。

核心思路是保留原始mask的位置信息，通过位置而非值来判断是否需要高亮。

方法一：按列传递mask信息 #

import pandas as pd
import numpy as np

dummy={'Location': {0: 'Perth', 1: 'Perth', 2: 'Perth', 3: 'Perth', 4: 'Perth', 5: 'Perth', 6: 'Perth', 7: 'Perth', 8: 'Perth', 9: 'Perth', 10: 'Perth', 11: 'Perth', 12: 'Perth', 13: 'Perth', 14: 'Perth', 15: 'Perth', 16: 'Perth', 17: 'Perth'}, 'Date': {0: '11/01/2012 0:00', 1: '11/01/2012 0:00', 2: '20/03/2012 0:00', 3: '6/06/2012 0:00', 4: '14/09/2012 0:00', 5: '17/12/2013 0:00', 6: '1/02/2014 0:00', 7: '1/02/2014 0:00', 8: '1/02/2014 0:00', 9: '1/02/2014 0:00', 10: '1/02/2014 0:00', 11: '1/02/2014 0:00', 12: '1/02/2014 0:00', 13: '1/02/2014 0:00', 14: '1/02/2014 0:00', 15: '1/02/2014 0:00', 16: '1/02/2014 0:00', 17: '1/02/2014 0:00'}, 'As µg/L': {0: '9630', 1: '9630', 2: '8580', 3: '4990', 4: '6100', 5: '282', 6: '21', 7: '<1', 8: '<1', 9: '<1', 10: '<1', 11: '<1', 12: '<1', 13: '<1', 14: '<1', 15: '<1', 16: '<1', 17: '<1'}, 'As': {0: '9.63', 1: '9.63', 2: '8.58', 3: '4.99', 4: '6.1', 5: '0.282', 6: '0.021', 7: '<1', 8: '<1', 9: '<1', 10: '<1', 11: '<1', 12: '<1', 13: '<1', 14: '<1', 15: '<1', 16: '<1', 17: '10'}, 'Ba': {0: 1000.0, 1: np.nan, 2: np.nan, 3: np.nan, 4: np.nan, 5: np.nan, 6: np.nan, 7: np.nan, 8: np.nan, 9: np.nan, 10: np.nan, 11: np.nan, 12: np.nan, 13: np.nan, 14: np.nan, 15: np.nan, 16: np.nan, 17: np.nan}, 'HCO3': {0: '10.00', 1: '0.50', 2: '0.50', 3: '<22', 4: '0.50', 5: '0.50', 6: '0.50', 7: np.nan, 8: np.nan, 9: np.nan, 10: '0.50', 11: np.nan, 12: np.nan, 13: np.nan, 14: np.nan, 15: np.nan, 16: np.nan, 17: np.nan}, 'Cd': {0: 0.0094, 1: 0.0094, 2: 0.011, 3: 0.0035, 4: 0.004, 5: 0.002, 6: 0.0019, 7: np.nan, 8: np.nan, 9: np.nan, 10: np.nan, 11: np.nan, 12: np.nan, 13: np.nan, 14: np.nan, 15: np.nan, 16: np.nan, 17: np.nan}, 'Ca': {0: 248.0, 1: 248.0, 2: 232.0, 3: 108.0, 4: 150.0, 5: 396.0, 6: 472.0, 7: np.nan, 8: np.nan, 9: np.nan, 10: np.nan, 11: np.nan, 12: np.nan, 13: np.nan, 14: 472.0, 15: np.nan, 16: np.nan, 17: np.nan}, 'CO3': {0: 0.5, 1: 0.5, 2: 0.5, 3: 0.5, 4: 0.5, 5: 0.5, 6: 0.5, 7: np.nan, 8: np.nan, 9: 0.5, 10: np.nan, 11: np.nan, 12: np.nan, 13: np.nan, 14: np.nan, 15: np.nan, 16: np.nan, 17: np.nan}, 'Cl': {0: 2.0, 1: 2.0, 2: 2.0, 3: 2.0, 4: 0.5, 5: 2.0, 6: 5.0, 7: np.nan, 8: np.nan, 9: np.nan, 10: np.nan, 11: np.nan, 12: np.nan, 13: 5.0, 14: np.nan, 15: np.nan, 16: np.nan, 17: np.nan}}

df=pd.DataFrame(dummy)
# 提前生成原始mask，保存位置信息
original_mask = df.applymap(lambda x: isinstance(x, str) and x.startswith('<'))

def remove_less_thans(x):
    if isinstance(x, (int, float)):
        return x
    elif isinstance(x, str) and x[0] == "<":
        try:
            return float(x[1:])/2
        except:
            return x
    elif isinstance(x, str):
        try:
            return float(x)
        except:
            return x
    else:
        return x

# 修改数据
df = df.applymap(remove_less_thans)

# 基于位置的样式函数
def highlight_modified(col):
    # 获取当前列