在两个数据框差异之间应用 Pandas 样式

Question

I've seen a lot of questions about finding the differences between two pandas dataframes, however here i'm trying to apply a Pandas.Style difference between two dataframes. Given these two example dataframes i'm hoping to have a formatter applied to right[1, "B"] and right["D"] because they're different that lefts values or new in general:

left = pd.DataFrame([[1,1,1], [2,2,2]], columns=list("ABC"))
right = pd.DataFrame([[1,1,10], [2,5,10]], columns=list("ABD"))

Here's my idea for the formatting method guided by the pandas documentation

def formatter(s, new):
    if s.name not in new.columns:
        # column doesn't exist strike through entire thing
        return "color: red; text-decoration: line-through;"

    elif not s.equals(new[s.name]):
        # apply per value a comparision of the elements
        # for val in s: 
            # if val != right[val.index??]:
                return "color: red; text-decoration: line-through;"

    return "color: black;"

left.style.apply(formatter, args=(right))

My thinking is that afterwards I should have something that is html like such:

 <table border="1" class="dataframe"> <thead> <tr style="text-align: right;"> <th></th> <th>A</th> <th>B</th> <th>C</th> </tr> </thead> <tbody> <tr> <th>0</th> <td>1</td> <td>1</td> <td>1</td> </tr> <tr> <th>1</th> <td>2</td> <td>2</td> <td>2</td> </tr> </tbody> </table>

 <table border="1" class="dataframe"> <thead> <tr style="text-align: right;"> <th></th> <th>A</th> <th>B</th> <th>C</th> </tr> </thead> <tbody> <tr> <th>0</th> <td>1</td> <td>1</td> <td style="color: red; text-decoration: line-through;">10</td> </tr> <tr> <th>1</th> <td>2</td> <td style="color: red; text-decoration: line-through;">5</td> <td style="color: red; text-decoration: line-through;">10</td> </tr> </tbody> </table>

Answer 1

It is a bit unclear exactly where you are a stuck, but the code is not far off.

This might be what you are after:

left = pd.DataFrame([[1,1,1], [2,2,2]], columns=list("ABC"))
right = pd.DataFrame([[1,1,10], [2,5,10]], columns=list("ABD"))
def formatter(s, new):
    if s.name not in new.columns:
        # column doesn't exist strike through entire thing
        return ["color: red; text-decoration: line-through;"]*len(s)

    elif not s.equals(new[s.name]):
        return ["color: red; text-decoration: line-through;" if v else "" for v in s == new[s.name]]

    return ["color: black;"]*len(s)

left.style.apply(formatter, args=[right])

The formatter method now returns data in the same shape as the input (as per the docs).

The right dataframe is passed as a list and not a tuple.

Also changed the per value comparison to return color if they differ, otherwise keep default style.