堆栈内存溢出
  简体   繁体   English

Pandas pivot_table：过滤聚合 function

[英]Pandas pivot_table: filter on aggregate function

 原文  2021-06-06 19:16:33  5   2    python/ pandas/ dataframe/ pandas-groupby/ pivot-table

问题描述

I am trying to pass a criteria to the aggregate function to pandas pivot_table and I am not able to figure out how to pass the criteria to the aggfunc.我正在尝试将标准传递给聚合 function 到 pandas pivot_table，但我无法弄清楚如何将标准传递给 aggfunc。 I have a data table which is converted to pandas df.我有一个转换为 pandas df 的数据表。

The input table data:输入表数据：

col1 col1 col2 col2 col3 col3 col4 col4 col5 col5 col6 col6 col7 col7
1 1 test1测试1 t1 t1 Dummy1假人1 result1结果1 10 10 102.2 102.2
2 2 test1测试1 t1 t1 Dummy2假人2 result2结果2 20 20 101.2 101.2
3 3 test1测试1 t1 t1 Dummy3假人3 result3结果3 30 30 102.3 102.3
4 4 test1测试1 t1 t1 Dummy4假人4 result4结果4 40 40 101.4 101.4
5 5 test2测试2 t1 t1 Dummy1假人1 result1结果1 10 10 100 100
6 6 test2测试2 t1 t1 Dummy2假人2 result2结果2 20 20 103 103
7 7 test2测试2 t1 t1 Dummy3假人3 result3结果3 30 30 104 104
8 8 test2测试2 t1 t1 Dummy4假人4 result4结果4 40 40 105 105
9 9 test3测试3 t1 t1 Dummy1假人1 result1结果1 10 10 102 102
10 10 test3测试3 t1 t1 Dummy2假人2 result2结果2 20 20 87 87
11 11 test3测试3 t1 t1 Dummy3假人3 result3结果3 30 30 107 107
12 12 test3测试3 t1 t1 Dummy5假人5 result4结果4 50 50 110.2 110.2
13 13 test4测试4 t1 t1 Dummy2假人2 result2结果2 20 20 120 120
14 14 test5测试5 t1 t1 Dummy6假人6 result1结果1 100 100 88 88
15 15 test1测试1 t1 t1 Dummy1假人1 result2结果2 10 10 106.2 106.2
16 16 test1测试1 t1 t1 Dummy1假人1 result6结果6 10 10 101.1 101.1

I want to get the maximum on col7 data, but only when the maximum is greater than 100. If any of the col7 data is greater than the user defined criteria, then all the other columns data needs to be populated irrespective if the data met the criteria or not.我想获得 col7 数据的最大值，但仅当最大值大于 100 时。如果任何 col7 数据大于用户定义的标准，则需要填充所有其他列数据，无论数据是否满足标准与否。

Tried the following:尝试了以下方法：

columns = ['col1', 'col2', 'col3', 'col4', 'col5', 'col6', 'col7']

df = pd.DataFrame({
    'col1': [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16],
    'col2': ['test1', 'test1', 'test1', 'test1', 'test2', 'test2', 'test2',
             'test2', 'test3', 'test3', 'test3', 'test3', 'test4', 'test5',
             'test1', 'test1'],
    'col3': ['t1', 't1', 't1', 't1', 't1', 't1', 't1', 't1', 't1', 't1', 't1',
             't1', 't1', 't1', 't1', 't1'],
    'col4': ['Dummy1', 'Dummy2', 'Dummy3', 'Dummy4', 'Dummy1', 'Dummy2',
             'Dummy3', 'Dummy4', 'Dummy1', 'Dummy2', 'Dummy3', 'Dummy5',
             'Dummy2', 'Dummy6', 'Dummy1', 'Dummy1'],
    'col5': ['result1', 'result2', 'result3', 'result4', 'result1', 'result2',
             'result3', 'result4', 'result1', 'result2', 'result3', 'result4',
             'result2', 'result1', 'result2', 'result6'],
    'col6': [10, 20, 30, 40, 10, 20, 30, 40, 10, 20, 30, 50, 20, 100, 10, 10],
    'col7': [100.2, 101.2, 102.3, 101.4, 100.0, 103.0, 104.0, 105.0, 102.0,
             87.0, 107.0, 110.2, 120.0, 88.0, 106.2, 101.1]
})

res=df.pivot_table(values = 'col7', index = ['col4', 'col5', 'col6'], columns = ['col2'], fill_value = '', aggfunc = 'max' >= 100)

TypeError: '>=' not supported between instances of 'str' and 'int'

Output should look like: Output 应如下所示：

Max pivoted output without col5:不带 col5 的最大旋转 output：

col4 col4 col6 col6 test1测试1 test2测试2 test3测试3 test4测试4 test5测试5
Dummy1假人1 10 10 106.2 106.2 100 100 102 102 N/A不适用 N/A不适用
Dummy2假人2 20 20 101.2 101.2 103 103 87 87 120 120 N/A不适用
Dummy3假人3 30 30 102.3 102.3 104 104 107 107 N/A不适用 N/A不适用
Dummy4假人4 40 40 101.4 101.4 105 105 N/A不适用 N/A不适用 N/A不适用
Dummy5假人5 50 50 N/A不适用 N/A不适用 110.2 110.2 N/A不适用 N/A不适用

Max pivoted output including col5:最大旋转 output 包括 col5：

col4 col4 col5 col5 col6 col6 test1测试1 test2测试2 test3测试3 test4测试4 test5测试5
Dummy1假人1 result2结果2 10 10 106.2 106.2 N/A不适用 N/A不适用 N/A不适用 N/A不适用
Dummy1假人1 result1结果1 10 10 102.2 102.2 100 100 102 102 N/A不适用 N/A不适用
Dummy2假人2 result2结果2 20 20 101.2 101.2 103 103 87 87 120 120 N/A不适用
Dummy3假人3 result3结果3 30 30 102.3 102.3 104 104 107 107 N/A不适用 N/A不适用
Dummy4假人4 result4结果4 40 40 101.4 101.4 105 105 N/A不适用 N/A不适用 N/A不适用
Dummy5假人5 result4结果4 50 50 N/A不适用 N/A不适用 110.2 110.2 N/A不适用 N/A不适用

Any guidance is much appreciated.非常感谢任何指导。

Thanks谢谢

2 个解决方案

解决方案1
 1 已采纳  2021-06-06 19:30:27

You can't compare the word 'max' to 100 via >= ( aggfunc = 'max' >= 100 ):您无法通过>= ( aggfunc = 'max' >= 100 ) 将单词 'max' 与 100 进行比较：

I recommend not setting the fill value to a string, masking the DataFrame, to get rid of undesired rows, then replace with empty string via fillna :我建议不要将填充值设置为字符串，屏蔽 DataFrame，以消除不需要的行，然后通过fillna替换为空字符串：

columns = ['col1', 'col2', 'col3', 'col4', 'col5', 'col6', 'col7']
res = df.pivot_table(values='col7', index=['col4', 'col5', 'col6'],
                     columns=['col2'], aggfunc='max')

col2                 test1  test2  test3  test4  test5
col4   col5    col6                                   
Dummy1 result1 10    102.2  100.0  102.0    NaN    NaN
       result2 10    106.2    NaN    NaN    NaN    NaN
       result6 10    101.1    NaN    NaN    NaN    NaN
Dummy2 result2 20    101.2  103.0   87.0  120.0    NaN
Dummy3 result3 30    102.3  104.0  107.0    NaN    NaN
Dummy4 result4 40    101.4  105.0    NaN    NaN    NaN
Dummy5 result4 50      NaN    NaN  110.2    NaN    NaN
Dummy6 result1 100     NaN    NaN    NaN    NaN   88.0

Mask for any values where any res >= 100 and fillna : any res >= 100和fillna的任何值的掩码：

res = res[(res >= 100).any(1)].fillna('')

col2                 test1  test2  test3  test4 test5
col4   col5    col6                                  
Dummy1 result1 10    102.2  100.0  102.0             
       result2 10    106.2                           
       result6 10    101.1                           
Dummy2 result2 20    101.2  103.0   87.0  120.0      
Dummy3 result3 30    102.3  104.0  107.0             
Dummy4 result4 40    101.4  105.0                    
Dummy5 result4 50                  110.2

Optional reset_index to clear the MultiIndex and rename_axis to clear the axis name:可选的reset_index清除 MultiIndex 和rename_axis清除轴名称：

res[(res >= 100).any(1)].fillna('').reset_index().rename_axis(None, axis=1)

     col4     col5  col6  test1  test2  test3  test4 test5
0  Dummy1  result1    10  102.2  100.0  102.0             
1  Dummy1  result2    10  106.2                           
2  Dummy1  result6    10  101.1                           
3  Dummy2  result2    20  101.2  103.0   87.0  120.0      
4  Dummy3  result3    30  102.3  104.0  107.0             
5  Dummy4  result4    40  101.4  105.0                    
6  Dummy5  result4    50                110.2

Complete Working Example:完整的工作示例：

import pandas as pd

df = pd.DataFrame({
    'col1': [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16],
    'col2': ['test1', 'test1', 'test1', 'test1', 'test2', 'test2', 'test2',
             'test2', 'test3', 'test3', 'test3', 'test3', 'test4', 'test5',
             'test1', 'test1'],
    'col3': ['t1', 't1', 't1', 't1', 't1', 't1', 't1', 't1', 't1', 't1', 't1',
             't1', 't1', 't1', 't1', 't1'],
    'col4': ['Dummy1', 'Dummy2', 'Dummy3', 'Dummy4', 'Dummy1', 'Dummy2',
             'Dummy3', 'Dummy4', 'Dummy1', 'Dummy2', 'Dummy3', 'Dummy5',
             'Dummy2', 'Dummy6', 'Dummy1', 'Dummy1'],
    'col5': ['result1', 'result2', 'result3', 'result4', 'result1', 'result2',
             'result3', 'result4', 'result1', 'result2', 'result3', 'result4',
             'result2', 'result1', 'result2', 'result6'],
    'col6': [10, 20, 30, 40, 10, 20, 30, 40, 10, 20, 30, 50, 20, 100, 10, 10],
    'col7': [102.2, 101.2, 102.3, 101.4, 100.0, 103.0, 104.0, 105.0, 102.0,
             87.0, 107.0, 110.2, 120.0, 88.0, 106.2, 101.1]
})

columns = ['col1', 'col2', 'col3', 'col4', 'col5', 'col6', 'col7']
res = df.pivot_table(values='col7', index=['col4', 'col5', 'col6'],
                     columns=['col2'], aggfunc='max')
res = (
    res[(res >= 100).any(1)].fillna('').reset_index().rename_axis(None, axis=1)
)
print(res)

To get the value without col5 remove it from the index of the pivot_table :要获取不带 col5 的值，请将其从pivot_table的index中删除：

columns = ['col1', 'col2', 'col3', 'col4', 'col5', 'col6', 'col7']
res = df.pivot_table(values='col7', index=['col4', 'col6'],
                     columns=['col2'], aggfunc='max')
res = (
    res[(res >= 100).any(1)].fillna('').reset_index().rename_axis(None, axis=1)
)

     col4  col6  test1  test2  test3  test4 test5
0  Dummy1    10  106.2  100.0  102.0             
1  Dummy2    20  101.2  103.0   87.0  120.0      
2  Dummy3    30  102.3  104.0  107.0             
3  Dummy4    40  101.4  105.0                    
4  Dummy5    50                110.2

解决方案2
 0  2021-06-06 19:48:40

Or you can try:或者您可以尝试：

res = df.assign(col7 = df.col7.where(df.col7 > 100)).pivot_table(values='col7', index=['col4', 'col5', 'col6'],
                     columns=['col2'], aggfunc='max', fill_value= '')

暂无
暂无

声明:本站的技术帖子网页，遵循CC BY-SA 4.0协议，如果您需要转载，请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.

相关问题 Pandas - 使用 pivot_table 聚合多个列 - Pandas - aggregate multiple columns with pivot_table Pandas pivot_table 中的函数参数 - Function arguments within Pandas pivot_table 重新排序 pandas pivot_table function 中的数据 - reorder data in pandas pivot_table function 熊猫Pivot_Table定义的函数aggfunc - Pandas Pivot_Table defined function aggfunc Pandas：数据透视表 - Pandas: pivot_table Pandas pivot_table - Pandas pivot_table pandas 在 pivot_table 期间聚合多个列 - pandas aggregate multiple columns during pivot_table 如何使用 pandas pivot_table 聚合唯一计数 - How to aggregate unique count with pandas pivot_table 将 Pandas pivot_table function 转换为 Polars pivot Function - Convert Pandas pivot_table function into Polars pivot Function 像常规数据框一样过滤熊猫pivot_table结果 - Filter pandas pivot_table result like a regular dataframe
相关标签
 
粤ICP备18138465号  © 2020-2024 STACKOOM.COM