Question

我有一些数据帧，使用以下代码生成：

from collections import defaultdict
import pandas as pd

mydict = { ('x305', 'BoxType1-1'): { 'box': 'x305', 'box#': '0', 'boxCode': 'Z8', 'version': '00.00' },
           ('x305', 'BoxType1-2'): { 'box': 'x305', 'box#': '0', 'boxCode': 'K8', 'version': '01.00' },
           ('x307', 'BoxType1-1'): { 'box': 'x307', 'box#': '0', 'boxCode': 'Z8', 'serialNo': 'None', 'version': '00.00' },
           ('x307', 'BoxType2-1'): { 'box': 'x307', 'box#': '0', 'boxCode': 'Z8', 'serialNo': 'None', 'version': '00.00' }, 
           ('x403', 'BoxType1-1'): { 'box': 'x403', 'box#': '0', 'boxCode': 'Z8', 'bla': 'None', 'version': '00.00' }, 
           ('x405', 'BoxType1-2'): { 'box': 'x405', 'box#': '0', 'boxCode': 'Z8', 'serialNo': 'None', 'version': '00.00' }, 
           ('x405', 'BoxType2-1'): { 'box': 'x405', 'box#': '0', 'boxCode': 'Z8', 'version': '00.00' }, 
           ('x510', 'BoxType1-3'): { 'box': 'x510', 'box#': '0', 'boxCode': 'Z8', 'version': '01.00' } }

boxTypes = [ 'BoxType1', 'BoxType2' ]
dataframes = defaultdict( set )

for boxType in boxTypes:
   dataframes[ boxType ] = pd.DataFrame.from_dict( { ( box, bt ): mydict[ ( box, bt ) ]
                                                     for box, bt in mydict.keys()
                                                     if boxType in bt },
                                                   orient='index' )

print dataframes[ 'BoxType1' ]
                 box   version box# boxCode  bla serialNo
x305 BoxType1-1  x305   00.00    0      Z8   NaN      NaN
     BoxType1-2  x305   01.00    0      K8   NaN      NaN
x307 BoxType1-1  x307   00.00    0      Z8   NaN     None
x403 BoxType1-1  x403   00.00    0      Z8  None      NaN
x405 BoxType1-2  x405   00.00    0      Z8   NaN     None
x510 BoxType1-3  x510   01.00    0      Z8   NaN      NaN

现在我试图找到计算整个数据帧中特定行元组的方法。例如，我想要一个这样的函数：

def countRowTuples( df, columns ):
    '''
    Count occurrences of row tuple in dataframe
    and return a new dataframe with a count column at the end
    '''

df2 = countRowTuples( dataframes['BoxType1'], columns=[ boxCode, bla, version ] )

df2 = 
                 box   version box# boxCode  bla serialNo  count
x305 BoxType1-1  x305   00.00    0      Z8   NaN      NaN    3
     BoxType1-2  x305   01.00    0      K8   NaN      NaN    1
x307 BoxType1-1  x307   00.00    0      Z8   NaN     None    3
x403 BoxType1-1  x403   00.00    0      Z8  None      NaN    1
x405 BoxType1-2  x405   00.00    0      Z8   NaN     None    3
x510 BoxType1-3  x510   01.00    0      Z8   NaN      NaN    1

或者，该函数可以摆脱原始索引和重复行，并返回如下数据框：

df2 = 
    version boxCode  bla   count
1   00.00     Z8     NaN     3
2   01.00     K8     NaN     1
3   00.00     Z8     None    1
4   01.00     Z8     NaN     1

任何人都有什么好的见解我会如何解决这个问题？

我尝试过以下操作，但一直返回一个空数据框:(。

df = dataframes[ 'BoxType1' ]
print df.groupby(df.columns.tolist()).size().reset_index().rename(columns={0:'count'})

Empty DataFrame
Columns: [box, version, box#, boxCode, bla, serialNo, count]
Index: []

Answer 1

由于空值，group by未执行所需的计数。试试这个：

dataframes['BoxType1'].fillna("NaN").groupby(["version", "boxCode", "bla"]).size().reset_index(name="count")

Answer 2

一种方法是添加另一个填充了一列的列，按照您想要不同值的字段分组（您需要用某些值填充这些NaN）并对添加的列求和。

$menus = get_registered_nav_menus();
foreach ( $menus as $location => $description ) {
    wp_nav_menu( array(
       'theme_location' => $location,
       'depth' => 1,
       'container' => 'div',
       'container_class' => 'col-md px-0 mb-5',
       'container_id' => $location,
       'menu_class' => 'navbar-nav h5',
       'fallback_cb' => 'WP_Bootstrap_Navwalker::fallback',
       'walker' => new WP_Bootstrap_Navwalker())
    );
}

如何计算数据框中的特定行元组？

2 个答案: