BUG: DataFrame.reindex not following limit - 方法参数的错误

[french-home]

Python版本: 3.6.5
版本:pandas == 0.24.2

import pandas as pd

Data = [
    ["A", "A", "A"],
    ["B", "B", "B"],
    ["C", "C", "C"],
    ["D", "D", "D"],
]
test1 = pd.DataFrame(Data)
print(test1)
print("------------")
test1 = test1.reindex([0, 1, 2, 3, 4, 5], method="ffill", limit=1)
print(test1)

limit参数限制数据继续向后填充

版本:pandas==0.25.1

import pandas as pd

Data = [
    ["A", "A", "A"],
    ["B", "B", "B"],
    ["C", "C", "C"],
    ["D", "D", "D"],
]
test1 = pd.DataFrame(Data)
print(test1)
print("------------")
test1 = test1.reindex([0, 1, 2, 3, 4, 5], method="ffill", limit=1)
print(test1)

limit参数并没有限制数据继续向后填充

[datapythonista]

谢谢 @french-home for reporting this. It'd be very useful if you can provide what's the output in both versions. Thanks!

[french-home]

Can I send it to you by mail?

…

------------------&nbsp;原始邮件&nbsp;------------------ 发件人:&nbsp;"Marc Garcia"<[email protected]&gt;; 发送时间:&nbsp;2019年9月26日(星期四) 晚上10:02 收件人:&nbsp;"pandas-dev/pandas"<[email protected]&gt;; 抄送:&nbsp;"French"<[email protected]&gt;;"Mention"<[email protected]&gt;; 主题:&nbsp;Re: [pandas-dev/pandas] BUG DataFream.rindex方法参数的错误 (#28631) 谢谢 @french-home for reporting this. It'd be very useful if you can provide what's the output in both versions. Thanks! — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or mute the thread.

[datapythonista]

Isn't it easier to simply copy it in a comment here? You can send it by email if that's a problem, but I think that can be easier.

[french-home]

Due to regional restrictions, it is difficult to access Github in China, so I can only send it to you by email.

…

------------------&nbsp;原始邮件&nbsp;------------------ 发件人:&nbsp;"Marc Garcia"<[email protected]&gt;; 发送时间:&nbsp;2019年9月26日(星期四) 晚上10:18 收件人:&nbsp;"pandas-dev/pandas"<[email protected]&gt;; 抄送:&nbsp;"French"<[email protected]&gt;;"Mention"<[email protected]&gt;; 主题:&nbsp;Re: [pandas-dev/pandas] BUG DataFream.rindex方法参数的错误 (#28631) Isn't it easier to simply copy it in a comment here? You can send it by email if that's a problem, but I think that can be easier. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or mute the thread.

[datapythonista]

Sure, please send it to ************* (edited) and I'll update the description for you.

[french-home]

I have sent the question to the email address you provided, please check it out!

…

------------------&nbsp;原始邮件&nbsp;------------------ 发件人:&nbsp;"Marc Garcia"<[email protected]&gt;; 发送时间:&nbsp;2019年9月26日(星期四) 晚上10:22 收件人:&nbsp;"pandas-dev/pandas"<[email protected]&gt;; 抄送:&nbsp;"French"<[email protected]&gt;;"Mention"<[email protected]&gt;; 主题:&nbsp;Re: [pandas-dev/pandas] BUG DataFream.rindex方法参数的错误 (#28631) Sure, please send it to [email protected] and I'll update the description for you. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or mute the thread.

[datapythonista]

Thanks a lot @french-home, it does look like a bug. I updated the description with your output, will have a look to see what's wrong when I have time.

Thanks for reporting it. If you'd like to research what's the problem yourself, and you need help with that, please let me know, happy to help.

[jorisvandenbossche]

So to make it a bit more explicit. Older versions correctly apply the "limit"

In [8]: pd.__version__  
Out[8]: '0.24.2'

In [9]: test1.reindex([0, 1, 2, 3, 4, 5], method="ffill", limit=1)
Out[9]: 
     0    1    2
0    A    A    A
1    B    B    B
2    C    C    C
3    D    D    D
4    D    D    D
5  NaN  NaN  NaN

while 0.25 / master does not:

In [22]: pd.__version__   
Out[22]: '0.26.0.dev0+418.g1cfba0a87'

In [23]: test1.reindex([0, 1, 2, 3, 4, 5], method="ffill", limit=1)
Out[23]: 
   0  1  2
0  A  A  A
1  B  B  B
2  C  C  C
3  D  D  D
4  D  D  D
5  D  D  D

So this is a regression.

[jorisvandenbossche]

So this boils down to RangeIndex.get_indexer no longer taking the limit keyword into account:

In [42]: idx = pd.Index(range(4))                                                                                                                                                                                  

In [43]: target = pd.Index([0, 1, 2, 3, 4, 5])                                                                                                                                                                     

In [44]: idx.get_indexer(target, method='pad', limit=1)                                                                                                                                                            
Out[44]: array([0, 1, 2, 3, 3, 3])

while the last element in that result should be -1.

This seems to be cause by #27119 (cc @toobaz )

[toobaz]

This seems to be cause by #27119 (cc @toobaz )

Definitely. I forgot to forward the limit argument here:

pandas/pandas/core/indexes/range.py

Line 354 in e9f6f6b

return super().get_indexer(target, method=method,