Question

我在查询Mongo-DB集合时遇到了困难。

我正在使用的文件

{
    CustomerId: 2,
    Addresses: [
        {
            Name: "Daniel"
        },
        {
            Name: "Eric"
        },
        {
            Name: "Dan"
        }
        ]
}

一些样本数据

var mdb = mongoClient.GetDatabase("test");
var collection = mdb.GetCollection<WebApi>("Customer");
var builder = Builders<Customer>.Filter;
var searchTerm = "Dan";

我现在想要通过CustomerId查询文档并过滤一些Addresses-values以返回名称为％Dan％的所有地址。由于地址集合可能很大，我想在查询时过滤这些地址

var filter = builder.And(builder.Eq("InstallationId", 2),
                         builder.Regex("Addresses.Name", new BsonRegularExpression(".*" + searchTerm + ".*", "i")))
var result = collection.Find(filter).FirstOrDefault();

当查询时，它可以工作，但它包含所有地址：

[
    {
        Name: "Daniel"
    },
    {
        Name: "Dan"
    }
]

我想得到的是：

{
    CustomerId: 2,
    Addresses: [
        {
            Name: "Daniel"
        },
        {
            Name: "Dan"
        }
        ]
}

或

collection.AsQueryable().Where(x => x.CustomId == 2).SelectMany(x =>
                x.Addresses.Where(a => a.Name.IndexOf(searchTerm, StringComparison.InvariantCultureIgnoreCase) >= 0).ToList();

我还尝试了几种使用Aggregate / Match / Project的方法但无法使其工作。 “AsQueryable”也不起作用，因为没有在驱动程序上实现IndexOf。

select 
   fromdate,
   employeename,
   noofleaves,       
   substring(employeename,1,charindex(',',employeename)-2) as empname,        
   substring(employeename,(charindex(',',employeename))+1,(len(employeename)-(charindex(',',employeename)))) as leavetype
from tablename

已使用的版本：

MongoDB.Driver v 2.6.1（也是Bson）
.Net Framework 4.5.2

Answer 1

这应该让你前进：

var result = collection
    .Aggregate()
    .Match(c => c.CustomerId == 2)
    .Project(c => new
        {
            c.CustomerId,
            Addresses = c.Addresses.Where(a => a.Name.IndexOf(searchTerm) != -1)
        })
    .ToList();

驱动程序会将其转换为：

db.Customer.aggregate([{
    $match: { CustomerId: 2 }
}, {
    $project: {
        CustomerId: "$CustomerId",
        Addresses: {
            $filter: {
                input: "$Addresses",
                as: "a",
                cond: {
                    $ne: [ { $indexOfBytes: [ "$$a.Name", "Dan" ] }, -1 ]
                }
            }
        },
        _id: 0
    }
}])

对于不区分大小写的版本，将来可以选择在某个阶段使用$ regex（请参阅https://jira.mongodb.org/browse/SERVER-11947）。但是，您现在可以使用ToUpper()：

var result = collection
    .Aggregate()
    .Match(c => c.CustomerId == 2)
    .Project(c => new
        {
            c.CustomerId,
            Addresses = c.Addresses.Where(a => a.Name.ToUpper().IndexOf(searchTerm.ToUpper()) != -1)
        })
    .ToList();

MongoDb / C＃过滤并获取所有子文档

1 个答案: