MongoDB $cond 或 $switch 条件表达式中的正则表达式？

Question

对于这样的收藏

{ _id: 1, name: "novel_1", qty: 15}
{ _id: 2, name: "magazine_1", qty: 5}
{ _id: 3, name: "novel_2", qty: 5}
{ _id: 4, name: "guitar_1", qty: 10}
{ _id: 5, name: "violin_1", qty: 10}

我想以某种方式使用 $project 管道根据项目名称对项目进行分类。然后从中得到一个分组计数。

db.items.aggregate([
    {$project: {category: {
        $switch: {
            branches: [
                // use regex here to categorize the items by their name
                {case: {$in: ['$name', [/magazine/, /novel/]]},
                    then: 'book'},
                {case: {$in: ['$name', [/guitar/, /violin/]]},
                    then: 'instrument'}
            ],
            default: 'others'
        }
    }}},
    // get the group-by count based on the category
    {$group: {
        _id: {category: '$category'},
        count: {$sum: '$qty'}
    }}
]);

然而，MongoDB 似乎不支持 $project 管道中的正则表达式条件表达式。那么我们如何进行这种先变换后分组查询呢？我想一种方法是通过 MapReduce，但据说性能不是很好。特别是我在我的应用程序中使用 python，使用 MapReduce 会将 JS 代码和 python 代码缠在一起。

Answer 1

您不需要 MapReduce。您可以使用聚合框架来执行此操作。

另外注意，你不需要先$project你的文档，你可以将$switch表达式传递给_id

db.items.aggregate(
    [
       {
          "$group": {
             "_id": {
                "$switch": {
                   "branches": [
                      {
                         "case": {
                            "$or": [
                               {
                                  "$gt": [
                                     {
                                        "$indexOfCP": [
                                           "$name",
                                           "magazine"
                                        ]
                                     },
                                     -1
                                  ]
                               },
                               {
                                  "$gt": [
                                     {
                                        "$indexOfCP": [
                                           "$name",
                                           "novel"
                                        ]
                                     },
                                     -1
                                  ]
                               }
                            ]
                         },
                         "then": "book"
                      },
                      {
                         "case": {
                            "$or": [
                               {
                                  "$gt": [
                                     {
                                        "$indexOfCP": [
                                           "$name",
                                           "violin"
                                        ]
                                     },
                                     -1
                                  ]
                               },
                               {
                                  "$gt": [
                                     {
                                        "$indexOfCP": [
                                           "$name",
                                           "guitar"
                                        ]
                                     },
                                     -1
                                  ]
                               }
                            ]
                         },
                         "then": "instrument"
                      }
                   ],
                   "default": "others"
                }
             },
             "count":{"$sum": "$qty"}
          }
       }
    ]
)

---

db.items.aggregate(
    [
       {
          "$group": {
             "_id": {
                "$switch": {
                   "branches": [
                      {
                         "case": {
                            "$gt": [
                               {
                                  "$size": {
                                     "$setInterserction": [
                                        {
                                           "$split": [
                                              "$name",
                                              "-"
                                           ]
                                        },
                                        [
                                           "magazine",
                                           "novel"
                                        ]
                                     ]
                                  }
                               },
                               0
                            ]
                         },
                         "then": "book"
                      },
                      {
                         "case": {
                            "$gt": [
                               {
                                  "$size": {
                                     "$setInterserction": [
                                        {
                                           "$split": [
                                              "$name",
                                              "-"
                                           ]
                                        },
                                        [
                                           "guitar",
                                           "violin"
                                        ]
                                     ]
                                  }
                               },
                               0
                            ]
                         },
                         "then": "instrument"
                      }
                   ],
                   "default": "others"
                }
             },
             "count": {"$sum": "$qty"}
          }
       }
    ]
)