Question

stores = [[232, '2016-02-05 04:30:00', 'Test User', 1],
          [332, '2016-02-06 04:30:00', 'Test User', 2],
          [432, '2016-02-07 04:30:00', 'Test User', 3],
          [532, '2016-02-08 04:30:00', 'Test User', 4],
          [632, '2016-02-09 04:30:00', 'Test User', 5]]

visits = pd.DataFrame(data=stores, columns=['store', 'visit', 'auditor', 'scene'])
visits.set_index(['store', 'visit'], inplace=True)
scenes = [[1, 1551, 2],
          [5, 1661, 4]]

scenes = pd.DataFrame(data=scenes, columns=['scene', 'product', 'amount'])
scenes.set_index('scene', inplace=True)

store_with_products = pd.merge(visits, scenes, left_on='scene', right_index=True, how='right')

我得到的结果如下：

                             auditor  scene  product  amount
store visit                                                 
232   2016-02-05 04:30:00  Test User      1     1551       2
632   2016-02-09 04:30:00  Test User      5     1661       4

但我正在做right join为什么我没有得到用NAN填充的完整存储矩阵，其中相关数据缺少场景矩阵？我如何解决上述问题？

Answer 1

您需要左连接，而不是右连接。然后它工作：

                             auditor  scene  product  amount
store visit                                                 
232   2016-02-05 04:30:00  Test User      1   1551.0     2.0
332   2016-02-06 04:30:00  Test User      2      NaN     NaN
432   2016-02-07 04:30:00  Test User      3      NaN     NaN
532   2016-02-08 04:30:00  Test User      4      NaN     NaN
632   2016-02-09 04:30:00  Test User      5   1661.0     4.0

正确加入熊猫

1 个答案: