Python 合并行+；Groupby/apply函数

我有如下数据帧（dataframe1）：

我想创建一个新的dataframe，若for unique ID的结束值等于另一行的开始值，那个么它应该合并到一行中（ID必须在同一个组中）。 所以，我想要这样的东西（dataframe2）：

现在对我来说太复杂了。所以首先我想到了groupby和apply。我在每个组中将结束列移动了一个（向下），并检查开始值与结束移位值相等的次数（我也可以在以后的数据集分析中使用它，所以它并没有那么无用）。所以我写了一个函数：

def mygroup(df):
    
    is_continued = 0
    
    df['End'] = df['End'].shift(1)
    for index, row in df.iterrows():
        if (row['Start'] == row['End']):
            is_continued = is_continued + 1
            
    return is_continued

然后：

is_continued = dataframe.groupby(['ID']).apply(mygroup)

我以为它会给我4个ID1和4个ID2，但不是

ID
ID1    4
ID2    0
dtype: int64

所以我的问题是

我在groupby/apply上做错了什么

如何合并这些行以获得dataframe2

---

我对熊猫几乎没有经验，因此（1）我不能回答你的第一个问题，（2）至于你的第二个问题，这可能不是专家会做的。但我将其作为一种解决方案：

import pandas as pd


def combine_rows(df):
    id = []
    group = []
    start = []
    end = []
    prev_row = None
    for index, row in df.iterrows():
        if prev_row is None:
            prev_row = row
            prev_end = prev_row.End
            continue
        if row.ID == prev_row.ID and row.Group == prev_row.Group and row.Start == prev_end:
            prev_end = row.End
        else:
            id.append(prev_row.ID)
            group.append(prev_row.Group)
            start.append(prev_row.Start)
            end.append(prev_end)
            prev_row = row
            prev_end = prev_row.End
    if prev_row is not None:
        id.append(prev_row.ID)
        group.append(prev_row.Group)
        start.append(prev_row.Start)
        end.append(prev_end)
    return pd.DataFrame({"ID": id, "Group": group, "Start": start, "End": end})

        
df = pd.DataFrame({
    "ID": ['ID1', 'ID1', 'ID1', 'ID1', 'ID1', 'ID2', 'ID2', 'ID2', 'ID2', 'ID2'],
    "Group": ['A', 'A', 'A', 'B', 'B', 'A', 'A', 'B', 'B', 'B'],
    "Start": [1, 2, 3, 4, 5, 6, 7, 8, 9, 11],
    "End": [2, 3, 4, 5, 6, 7, 8, 9, 10, 12]
})
print(df)
df = combine_rows(df)
print(df))

印刷品：

    ID Group  Start  End
0  ID1     A      1    2
1  ID1     A      2    3
2  ID1     A      3    4
3  ID1     B      4    5
4  ID1     B      5    6
5  ID2     A      6    7
6  ID2     A      7    8
7  ID2     B      8    9
8  ID2     B      9   10
9  ID2     B     11   12
    ID Group  Start  End
0  ID1     A      1    4
1  ID1     B      4    6
2  ID2     A      6    8
3  ID2     B      8   10
4  ID2     B     11   12

import pandas as pd


def combine_rows(df):
    id = []
    group = []
    start = []
    end = []
    prev_row = None
    for index, row in df.iterrows():
        if prev_row is None:
            prev_row = row
            prev_end = prev_row.End
            continue
        if row.ID == prev_row.ID and row.Group == prev_row.Group and row.Start == prev_end:
            prev_end = row.End
        else:
            id.append(prev_row.ID)
            group.append(prev_row.Group)
            start.append(prev_row.Start)
            end.append(prev_end)
            prev_row = row
            prev_end = prev_row.End
    if prev_row is not None:
        id.append(prev_row.ID)
        group.append(prev_row.Group)
        start.append(prev_row.Start)
        end.append(prev_end)
    return pd.DataFrame({"ID": id, "Group": group, "Start": start, "End": end})

        
df = pd.DataFrame({
    "ID": ['ID1', 'ID1', 'ID1', 'ID1', 'ID1', 'ID2', 'ID2', 'ID2', 'ID2', 'ID2'],
    "Group": ['A', 'A', 'A', 'B', 'B', 'A', 'A', 'B', 'B', 'B'],
    "Start": [1, 2, 3, 4, 5, 6, 7, 8, 9, 11],
    "End": [2, 3, 4, 5, 6, 7, 8, 9, 10, 12]
})
print(df)
df = combine_rows(df)
print(df))

    ID Group  Start  End
0  ID1     A      1    2
1  ID1     A      2    3
2  ID1     A      3    4
3  ID1     B      4    5
4  ID1     B      5    6
5  ID2     A      6    7
6  ID2     A      7    8
7  ID2     B      8    9
8  ID2     B      9   10
9  ID2     B     11   12
    ID Group  Start  End
0  ID1     A      1    4
1  ID1     B      4    6
2  ID2     A      6    8
3  ID2     B      8   10
4  ID2     B     11   12