wrong aggregated results when pivot_table index is ordered categorical data

[peterpanmj]

Code Sample, a copy-pastable example if possible

import pandas as pd
import numpy as np

_c = np.arange(5,0,-1)

df = pd.DataFrame(np.stack([_c]*3).T, columns=['A', 'B', 'C'])

df["D"] = pd.Series(["A","B","B","A", "A"])

df["E"] = pd.cut(x=df["A"],bins=[0,3,6,10])

print(pd.pivot_table(df, index='E', columns=['D'], values="B", aggfunc="sum", margins=True))

Output is

D       A  B  All
E                
(0, 3]  3  3    9
(3, 6]  5  4    6
All     8  7   15

Problem description

the data is as follows:

	A	B	C	D	E
0	5	5	5	A	(3, 6]
1	4	4	4	B	(3, 6]
2	3	3	3	B	(0, 3]
3	2	2	2	A	(0, 3]
4	1	1	1	A	(0, 3]

Margin calculation results are in wrong order
(0, 3] should have a sum value of 6,
(3, 6] should have a sum value of 9

If I dont use column argument, results are correct.

B
E	
(0, 3]	6
(3, 6]	9
(6, 10]	0
All	15

I've tried both .23.4 and .24.2 . They have the same results.

Expected Output

D       A  B  All
E                
(0, 3]  3  3    6
(3, 6]  5  4    9
All     8  7   15

Output of pd.show_versions()

INSTALLED VERSIONS

commit: None
python: 3.7.1.final.0
python-bits: 64
OS: Windows
OS-release: 10
machine: AMD64
processor: Intel64 Family 6 Model 142 Stepping 9, GenuineIntel
byteorder: little
LC_ALL: None
LANG: zh_CN.GBK@cjknarrow
LOCALE: None.None

pandas: 0.23.4
pytest: 4.0.2
pip: 18.1
setuptools: 40.6.3
Cython: 0.29.2
numpy: 1.15.4
scipy: 1.1.0
pyarrow: None
xarray: None
IPython: 7.2.0
sphinx: 1.8.2
patsy: 0.5.1
dateutil: 2.7.5
pytz: 2018.7
blosc: None
bottleneck: 1.2.1
tables: 3.4.4
numexpr: 2.6.8
feather: None
matplotlib: 3.0.2
openpyxl: 2.5.12
xlrd: 1.2.0
xlwt: 1.3.0
xlsxwriter: 1.1.2
lxml: 4.2.5
bs4: 4.6.3
html5lib: 1.0.1
sqlalchemy: 1.2.15
pymysql: None
psycopg2: None
jinja2: 2.10
s3fs: None
fastparquet: None
pandas_gbq: None
pandas_datareader: None

[jreback]

check on 0.24.2 i believe we patched this

[peterpanmj]

#25814

I tried on 0.24.2. It gives me an error. It might be gone, but I cant tell.

check on 0.24.2 i believe we patched this

[peterpanmj]

In [19]: import pandas as pd
    ...: import numpy as np
    ...:
    ...: _c = np.arange(5,0,-1)
    ...:
    ...: df = pd.DataFrame(np.stack([_c]*3).T, columns=['A', 'B', 'C'])
    ...:
    ...: df["D"] = pd.Series(["A","B","B","A", "A"])
    ...:
    ...: df["E"] = pd.cut(x=df["A"],bins=[0,3,6,10]).cat.add_categories("Miss")
    ...:
    ...:
    ...:

In [20]: df
Out[20]:
   A  B  C  D       E
0  5  5  5  A  (3, 6]
1  4  4  4  B  (3, 6]
2  3  3  3  B  (0, 3]
3  2  2  2  A  (0, 3]
4  1  1  1  A  (0, 3]

In [21]: print(pd.pivot_table(df, index='E', columns=['D'], values="B", aggfunc="
    ...: sum", margins=True))
D       A  B  All
E
(0, 3]  3  3    9
(3, 6]  5  4    6
All     8  7   15

I can bypass 
#25814  , by adding a new category.  However the results are still wrong. The above code is tested on current master.
pandas: 0.25.0.dev0+299.ge31b4f46d

[mukundm19]

I would like to take a shot at this.

[peterpanmj]

@mukundm19 Any progress yet ?

[dotboris]

There's a dirty workaround for this. You could convert your intervals to strings. Obviously you lose all of the properties of the intervals. If you're OK with that (lets say you only show the intervals and don't operate on them afterwards) then you're golden.

You can workaround this by adding .astype(str) after the call to .cut(...).

import pandas as pd
import numpy as np

_c = np.arange(5,0,-1)

df = pd.DataFrame(np.stack([_c]*3).T, columns=['A', 'B', 'C'])

df["D"] = pd.Series(["A","B","B","A", "A"])

df["E"] = pd.cut(x=df["A"],bins=[0,3,6,10]).astype(str)

print(pd.pivot_table(df, index='E', columns=['D'], values="B", aggfunc="sum", margins=True))

D       A  B  All
E
(0, 3]  3  3    6
(3, 6]  5  4    9
All     8  7   15

[jschendel]

This looks to be fixed on master now, and just could use a test:

In [1]: import numpy as np; import pandas as pd; pd.__version__
Out[1]: '0.25.0.dev0+833.gad18ea35b'

In [2]: _c = np.arange(5,0,-1) 
   ...: df = pd.DataFrame(np.stack([_c]*3).T, columns=['A', 'B', 'C']) 
   ...: df["D"] = pd.Series(["A","B","B","A", "A"]) 
   ...: df["E"] = pd.cut(x=df["A"],bins=[0,3,6,10])

In [3]: df
Out[3]: 
   A  B  C  D       E
0  5  5  5  A  (3, 6]
1  4  4  4  B  (3, 6]
2  3  3  3  B  (0, 3]
3  2  2  2  A  (0, 3]
4  1  1  1  A  (0, 3]

In [4]: pd.pivot_table(df, index='E', columns=['D'], values="B", aggfunc="sum", margins=True)
Out[4]: 
D       A  B  All
E                
(0, 3]  3  3    6
(3, 6]  5  4    9
All     8  7   15