Python / pingouin :如果所有元素的x-y都为零，则ValueError: zero_method 'wilcox‘和'pratt’不起作用

Question

我得到了这个数据集，(mydata.csv)，我想做一个非参数配对测试(wilcoxon)，但当我运行代码时，出现了这个错误，我不明白它的含义和如何解决：

import pandas as pd 
import pingouin as pg

    df = pd.read_csv("mydata.csv") 

             id        price      review     score
    0       7949480    99.0       Check-in   10.0
    1       6627449    125.0      Check-in   10.0
    2       5557381    69.0       Check-in   10.0
    3       9147025    125.0      Check-in   10.0
    4       11675715   85.0       Check-in   10.0
    ... ... ... ... ...
    273745  12288416   130.0      Value      10.0
    273746  7930288    95.0       Value      10.0
    273747  18342528   50.0       Value      10.0
    273748  16232278   42.0       Value       9.0
    273749  18223756   115.0      Value      10.0

    pg.pairwise_ttests(dv="price", within="review", subject="id", data=df, parametric=False).round(3)

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-17-cb7090cfec05> in <module>
----> 1 pg.pairwise_ttests(dv="price", within="review", subject="id", data=reviews_test, parametric=False).round(3)

~\anaconda3\lib\site-packages\pingouin\pairwise.py in pairwise_ttests(data, dv, between, within, subject, parametric, marginal, alpha, tail, padjust, effsize, correction, nan_policy, return_desc, interaction, within_first)
    402                 if paired:
    403                     stat_name = 'W-val'
--> 404                     df_ttest = wilcoxon(x, y, tail=tail)
    405                 else:
    406                     stat_name = 'U-val'

~\anaconda3\lib\site-packages\pingouin\nonparametric.py in wilcoxon(x, y, tail)
    443 
    444     # Compute test
--> 445     wval, pval = scipy.stats.wilcoxon(x, y, zero_method='wilcox',
    446                                       correction=True, alternative=tail)
    447 

~\anaconda3\lib\site-packages\scipy\stats\morestats.py in wilcoxon(x, y, zero_method, correction, alternative, mode)
   2961         if zero_method in ["wilcox", "pratt"]:
   2962             if n_zero == len(d):
-> 2963                 raise ValueError("zero_method 'wilcox' and 'pratt' do not "
   2964                                  "work if x - y is zero for all elements.")
   2965         if zero_method == "wilcox":

ValueError: zero_method 'wilcox' and 'pratt' do not work if x - y is zero for all elements.

Answer 1

输入注释太长，旋转有问题，在同一组下有相同的值：

import pandas as pd 
import pingouin as pg
df = pd.read_csv("https://github.com/felfonsecal/StackOverflowQuestions/raw/main/mydata.csv")

df[df['id']==7949480]

    id  price   review  score
0   7949480 99.0    Check-in    10.0
45625   7949480 99.0    Cleanliness 10.0
91250   7949480 99.0    Communication   10.0
136875  7949480 99.0    Location    10.0
182500  7949480 99.0    Rating  10.0
228125  7949480 99.0    Value   10.0

df.groupby(['id'])['price'].nunique()

id
590         1
592         1
686         1
930         1
1235        1
           ..
18561941    1
18562129    1
18569355    1
18577490    1
18598103    1

对于跨组的任何个人，在值没有变化的情况下测试这样的数据是完全没有意义的