Pandas - 如果在列表中找到重复行并计数答案

【问题标题】：Pandas - Repeat row if found in list and countPandas - 如果在列表中找到重复行并计数
【发布时间】：2022-08-09 14:12:48
【问题描述】：

如果在列表中找到重复行需要帮助。

如果在列表增量计数中找到值
如果找到多个实例，则通过递增计数重复该行

数据框：

data = {\'First_Name\':[\'Tom\', \'Nick\', \'Daniel\', \'Jack\'],
        \'oter_col\':[\'other_value1\', \'other_value2\', \'other_value3\', \'other_value4\']}
df = pd.DataFrame(data)
df[\"repeation\"] = 0

Df 看起来像：

  First_Name      oter_col  count
0        Tom  other_value1      0
1       Nick  other_value2      0
2     Daniel  other_value3      0
3       Jack  other_value4      0

输入列表：

Full_Name = [\"Tom Cruise\", \"Tom Holland\", \"Tom Hardy\", \"Jack black\", \"Chris Hemsworth\"]

需要输出如：

  First_Name    Full_Name      oter_col  count
0        Tom   Tom Cruise  other_value1      1
1        Tom  Tom Holland  other_value1      2
2        Tom    Tom Hardy  other_value1      3
3       Jack   Jack black  other_value4      1
4       Nick               other_value2      1
5     Daniel               other_value3      0

尝试过这样的事情来获得第一个匹配索引，所以我可以重复该行但不知道该怎么做。

for name in Full_Name:
    m = df.First_Name.str.contains(name.split()[0])
    first_index_found = m.idxmax() if m.any() else None
    if type(first_index_found) == int:
        print(first_index_found)

标签： python pandas dataframe

【解决方案1】：

您可以使用：

# Create a regex pattern to extract First_Name from Full_Name
pat = fr"\b({'|'.join(df['First_Name'])})\b"

# Create a dataframe from Full_Name
df1 = pd.DataFrame({'Full_Name': Full_Name})
df1['First_Name'] = df1['Full_Name'].str.extract(pat)

# Merge them on First_Name column
out = df.merge(df1, on='First_Name', how='left')

# Count (choose one)
out['repeation'] = out.groupby('First_Name').cumcount().add(1)
# OR
out['repeation2'] = (out.dropna().groupby('First_Name').cumcount().add(1)
                        .reindex(out.index, fill_value=0))

输出：

>>> out
  First_Name      oter_col  repeation    Full_Name  repeation2
0        Tom  other_value1          1   Tom Cruise           1
1        Tom  other_value1          2  Tom Holland           2
2        Tom  other_value1          3    Tom Hardy           3
3       Nick  other_value2          1          NaN           0
4     Daniel  other_value3          1          NaN           0
5       Jack  other_value4          1   Jack black           1

【讨论】：

【解决方案2】：

您可以将 DataFrame 制作为 merge 并将非 NA 名称计算为每组 cumsum：

df2 = (pd.DataFrame({'Full_Name': Full_Name})
         .assign(First_Name=lambda d: d['Full_Name'].str.extract('(\w+)'))
      )
(df
 .merge(df2, on='First_Name', how='left')
 .assign(repeation=lambda d: d['Full_Name'].notna().groupby(d['First_Name']).cumsum())
)

输出：

  First_Name      oter_col  repeation    Full_Name
0        Tom  other_value1          1   Tom Cruise
1        Tom  other_value1          2  Tom Holland
2        Tom  other_value1          3    Tom Hardy
3       Nick  other_value2          0          NaN
4     Daniel  other_value3          0          NaN
5       Jack  other_value4          1   Jack black

【讨论】：

我犹豫在 groupby 之前删除 NaN 但我认为你是对的。预期的输出不一致。