如何将列表的 for 循环替换为数据框列

Question

我有数据框df_paths，我想从列路径中的列表中创建列 path_pair。输出选择如下表所示：

session_id	path	path_pair
T01	[Start, play_series, Null]	[[Start, play_series],[ play_series, Null]]
T02	[Start, play_tv, purchase, Conversion]	[[Start, play_tv],[play_tv,purchase],[purchase, Conversion]]

我的脚本：

for index, row in df_paths.iterrows():
  list_pair = []
  print(index, row['path'])

  for i, val in enumerate(row['path']):
    if i != len(row['path'])-1:
      row_ls = [row['path'][i], row['path'][i+1]]
      # append row list to ls
      list_pair.append(row_ls)
  print(list_pair)

  row['path_pair'] = np.array(row['path'])
  list_new = np.array(list_pair)

  df_paths.at[index, 'path_pair'] = row['path_pair'][list_new] 
  print(df_paths.loc[index]['path_pair'])

错误： IndexError: arrays used as indices must be of integer (or boolean) type

Answer 1

data = [['T01', ['Start', 'play_series', 'Null']], [
    'T02', ['Start', 'play_tv', 'purchase', 'Conversion']]]
df = pd.DataFrame(data, columns=['session_id', 'path'])
df['path_pair'] = df['path'].apply(lambda x: list(zip(x, x[1:])))
df