文档菜单
文档首页
/
语言
/
C#/.NET
/
C#/.NET
/
基础知识

Atlas搜索

本页内容

  • 概述
  • 创建Atlas搜索索引
  • Atlas搜索运算符
  • 自动完成
  • 复合
  • 嵌入式文档
  • 等于
  • 存在
  • 地理形状
  • 地理包含
  • 类似内容
  • 附近
  • 短语
  • 查询字符串
  • 范围
  • 正则表达式
  • 跨度
  • 文本
  • 通配符

概述

在这份指南中，您可以了解如何使用搜索构建器使用MongoDB .NET/C#驱动程序构建$search聚合管道阶段。

要了解更多关于$search聚合管道阶段的信息，请参阅$search.

注意

仅在MongoDB v4.2及以上版本的Atlas上可用

$search聚合管道操作符仅适用于托管在MongoDB Atlas集群上的集合，这些集群运行MongoDB v4.2或更高版本，并且受Atlas搜索索引的保护。要了解更多关于所需设置和此操作符的功能，请参阅Atlas Search文档。

本指南中的示例使用名为guitars的集合中的以下文档。

{ "_id": 1, "make": "Fender", "description": "Classic guitars known for their versatility.", "establishedYear": 1946, "in_stock": true, "rating": 9 }
{ "_id": 2, "make": "Gibson", "description": "Classic guitars known for their rich, full tones.", "establishedYear": 1902, "in_stock": true, "rating": 8 }
{ "_id": 3, "make": "PRS", "description": "High-end guitars known for their quality.", "establishedYear": 1985, "in_stock": true, "rating": 9 }
{ "_id": 4, "make": "Kiesel", "description": "Quality guitars made only for custom orders.", "establishedYear": 2015, "in_stock": false }
{ "_id": 5, "make": "Ibanez", "description": "Well-crafted guitars used by many professional guitarists.", "establishedYear": 1957, "in_stock": true, "rating": 7 }
{ "_id": 6, "make": "Strandberg", "description": "Modern guitars known for their headless models.", "establishedYear": 1982, "in_stock": false }

以下Guitar类表示该集合中的文档。

public class Guitar
{
    public int Id { get; set; }
    public string Make { get; set; }
    public List<string> Models { get; set; }
    public int EstablishedYear { get; set; }
    [BsonElement("in_stock")]
    public bool InStock { get; set; }
    public int? Rating { get; set; }
}

注意

guitars集合中的文档使用驼峰命名约定。本指南中的示例使用ConventionPack将集合中的字段反序列化为Pascal大小写，并将它们映射到Guitar类的属性。

要了解更多关于自定义序列化的信息，请参阅自定义序列化.

创建Atlas搜索索引

在您可以对Atlas集合进行搜索之前，您必须首先在该集合上创建一个Atlas搜索索引。Atlas搜索索引是一种将数据分类为可搜索格式的数据结构。

有关创建Atlas搜索索引的说明，请参阅创建Atlas搜索索引 Atlas指南。

Atlas搜索操作符

Search类包含您可以使用的方法来执行$search操作。有关可用的$search操作符的完整列表，请参阅操作符和收集器 Atlas指南。

自动完成

使用Autocomplete()方法搜索包含来自不完整输入字符串的字符序列的单词或短语。

以下示例使用字符串 "Gib" 在 make 字段上对 guitars 集合进行自动完成搜索。

var result = guitarsCollection.Aggregate()
    .Search(Builders<Guitar>.Search.Autocomplete(g => g.Make, "Gib"), indexName: "guitarmakes")
    .ToList();

注意

如果您正在搜索的字段由搜索索引索引，您必须将索引名称传递给 Autocomplete 调用。如果不存在搜索索引，则使用默认索引。

搜索返回以下文档

{ "_id" : 2, "make" : "Gibson", "description" : "Classic guitars known for their rich, full tones.", "establishedYear" : 1902, "in_stock" : true, "rating" : 8 }

有关 autocomplete 操作符的更多信息，请参阅 autocomplete Atlas指南。

复合

使用 Compound() 方法将两个或多个操作符组合成单个搜索。

以下示例在 guitars 集合中搜索符合以下所有条件的文档

  • 文档上存在 rating 字段

  • in_stock 字段不是 false

  • establishedYear 字段具有大于 1940 的值

var result = guitarsCollection.Aggregate()
    .Search(Builders<Guitar>.Search.Compound()
        .Must(Builders<Guitar>.Search.Exists(g => g.Rating))
        .MustNot(Builders<Guitar>.Search.Equals(g => g.InStock, false))
        .Must(Builders<Guitar>.Search.Range(g => g.EstablishedYear, SearchRangeBuilder.Gt(1940))))
    .ToList();

搜索返回以下文档

{ "_id" : 1, "make" : "Fender", "description" : "...", "establishedYear" : 1946, "in_stock" : true, "rating" : 9 }
{ "_id" : 3, "make" : "PRS", "description" : "...", "establishedYear" : 1985, "in_stock" : true, "rating" : 9 }
{ "_id" : 5, "make" : "Ibanez", "description" : "...", "establishedYear" : 1957, "in_stock" : true, "rating" : 7 }

有关 compound 操作符的更多信息，请参阅 compound Atlas指南。

内嵌文档

使用 EmbeddedDocument() 方法对字段数组值中的文档执行搜索操作。

注意

要搜索内嵌文档，必须在数组字段上创建 embeddedDocument 索引。

要了解如何定义一个嵌入式文档索引，请参阅Atlas文档中的定义嵌入式文档类型的索引。

考虑到guitars集合中的一些文档添加了包含产品详情对象数组的productDetails字段。

{ "_id": 1, "make": "Fender", "description": "...", "establishedYear": 1946, "in_stock": true, "rating": 9, "productDetails": [{"product_id": 1234, "serial": "YZ5678"}] }
{ "_id": 2, "make": "Gibson", "description": "...", "establishedYear": 1902, "in_stock": true, "rating": 8 }
{ "_id": 3, "make": "PRS", "description": "...", "establishedYear": 1985, "in_stock": true, "rating": 9, "productDetails": [{"product_id": 9870, "serial": "AB5555"}] }
{ "_id": 4, "make": "Kiesel", "description": "...", "establishedYear": 2015, "in_stock": false }
{ "_id": 5, "make": "Ibanez", "description": "...", "establishedYear": 1957, "in_stock": true, "rating": 7, "productDetails": [{"product_id": 5432, "serial": "ZZ1234"}] }
{ "_id": 6, "make": "Strandberg", "description": "...", "establishedYear": 1982, "in_stock": false }

在productDetails字段上创建嵌入式文档索引后，您可以对该字段中的文档执行Atlas搜索操作。以下示例在productDetails数组字段上执行文本搜索，并返回任何具有serial字段值为"YZ5678"的文档。

var result = guitarsCollection.Aggregate()
    .Search(Builders<Guitar>.Search.EmbeddedDocument(
        g => g.ProductDetails,
        Builders<ProductDetails>.Search.Text(p => p.Serial, "YZ5678")
    )).ToList();
return result;

搜索返回以下文档

{ "_id" : 1, "make" : "Fender", "description" : "Classic guitars known for their versatility.", "establishedYear" : 1946, "in_stock" : true, "rating" : 9, "productDetails" : [{ "product_id" : 1234, "serial" : "YZ5678" }] }

要了解有关嵌入式文档运算符的更多信息，请参阅Atlas指南中的嵌入式文档。

等于

使用Equals()方法检查字段是否匹配指定值。

以下示例在guitars集合中搜索任何in_stock字段值为true的文档。

var result = guitarsCollection.Aggregate()
    .Search(Builders<Guitar>.Search.Equals(g => g.InStock, true))
    .ToList();

搜索返回以下文档

{ "_id" : 1, "make" : "Fender", "description" : "...", "establishedYear" : 1946, "in_stock" : true, "rating" : 9 }
{ "_id" : 2, "make" : "Gibson", "description" : "...", "establishedYear" : 1902, "in_stock" : true, "rating" : 8 }
{ "_id" : 3, "make" : "PRS", "description" : "...", "establishedYear" : 1985, "in_stock" : true, "rating" : 9 }
{ "_id" : 5, "make" : "Ibanez", "description" : "...", "establishedYear" : 1957, "in_stock" : true, "rating" : 7 }

要了解有关equals运算符的更多信息，请参阅Atlas指南中的equals。

存在

使用Exists()方法搜索具有指定索引字段名的文档。如果指定的字段存在但未索引，则该文档不包括在结果集中。

以下示例在guitars集合中搜索任何存在rating字段的文档。

var result = guitarsCollection.Aggregate()
    .Search(Builders<Guitar>.Search.Exists(g => g.Rating))
    .ToList();

搜索返回以下文档

{ "_id" : 1, "make" : "Fender", "description" : "...", "establishedYear" : 1946, "in_stock" : true, "rating" : 9 }
{ "_id" : 2, "make" : "Gibson", "description" : "...", "establishedYear" : 1902, "in_stock" : true, "rating" : 8 }
{ "_id" : 3, "make" : "PRS", "description" : "...", "establishedYear" : 1985, "in_stock" : true, "rating" : 9 }
{ "_id" : 5, "make" : "Ibanez", "description" : "...", "establishedYear" : 1957, "in_stock" : true, "rating" : 7 }

要了解有关exists运算符的更多信息，请参阅Atlas指南中的exists。

GeoShape

使用 GeoShape() 方法来搜索与给定几何形状相关的文档。在指定搜索坐标时，必须首先指定经度，然后是纬度。经度值可以是 -180 到 180，纬度值可以是 -90 到 90。

注意

Atlas Search 不支持以下内容

  • 非默认坐标系（CRS）

  • 平面 XY 坐标系（二维）

  • 坐标对点表示法（pointFieldName: [12, 34]）

假设在 guitars 集合中的一些文档添加了 in_stock_location 字段。集合中更改后的文档现在如下所示

{ "_id": 1, "make": "Fender", "description": "...", "establishedYear": 1946, "in_stock": true, "in_stock_location":  { "type": "Point", "coordinates": [ -73.93615, 40.69791 ]}, "rating": 9 }
{ "_id": 2, "make": "Gibson", "description": "...", "establishedYear": 1902, "in_stock": true, "in_stock_location":  { "type": "Point", "coordinates": [ 47.6062, 122.321 ]}, "rating": 8 }

以下示例搜索所有在 in_stock_location 字段中的坐标与指定的多边形相交的文档

GeoJsonPolygon<GeoJson2DGeographicCoordinates> searchArea = new(new(new(new GeoJson2DGeographicCoordinates[]
{
     new(-72.93615, 41.69791),
     new(-72.93615, 40.59791),
     new(-74.93615, 40.59791),
     new(-74.93615, 41.69791),
     new(-72.93615, 41.69791),
})));
var result = guitarsCollection.Aggregate()
    .Search(Builders<Guitar>.Search.GeoShape(g => g.InStockLocation, GeoShapeRelation.Intersects, searchArea))
    .ToList();

搜索返回以下文档

{ "_id" : 1, "make" : "Fender", "description" : "...", "establishedYear" : 1946, "in_stock" : true, "in_stock_location" : { "type" : "Point", "coordinates" : ["-73.93615", "40.69791"] }, "rating" : 9 }

有关 geoShape 操作符的更多信息，请参阅 geoShape Atlas 指南。

GeoWithin

使用 GeoWithin() 方法来搜索坐标在给定几何形状内的指定 GeoJSON 字段中的文档。您可以搜索在以下范围内的点

  • 圆形

  • 边界框

  • 多边形

在指定搜索坐标时，必须首先指定经度，然后是纬度。经度值可以是 -180 到 180，纬度值可以是 -90 到 90。

注意

Atlas Search 不支持以下内容

  • 非默认坐标系（CRS）

  • 平面 XY 坐标系（二维）

  • 坐标对点表示法（pointFieldName: [12, 34]）

假设在 guitars 集合中的一些文档添加了 in_stock_location 字段。集合中更改后的文档现在如下所示

{ "_id": 1, "make": "Fender", "description": "...", "establishedYear": 1946, "in_stock": true, "in_stock_location":  { "type": "Point", "coordinates": [ -73.93615, 40.69791 ]}, "rating": 9 }
{ "_id": 2, "make": "Gibson", "description": "...", "establishedYear": 1902, "in_stock": true, "in_stock_location":  { "type": "Point", "coordinates": [ 47.6062, 122.321 ]}, "rating": 8 }

以下示例搜索所有在 in_stock_location 字段中的坐标位于指定多边形内的文档

GeoJsonPolygon<GeoJson2DGeographicCoordinates> searchArea = new(new(new(new GeoJson2DGeographicCoordinates[]
{
     new(-74.3994140625, 40.5305017757),
     new(-74.7290039063, 40.5805846641),
     new(-74.7729492188, 40.9467136651),
     new(-74.0698242188, 41.1290213475),
     new(-73.65234375, 40.9964840144),
     new(-72.6416015625, 40.9467136651),
     new(-72.3559570313, 40.7971774152),
     new(-74.3994140625, 40.5305017757),
})));
var result = guitarsCollection.Aggregate()
    .Search(Builders<Guitar>.Search.GeoWithin(g => g.InStockLocation, searchArea))
    .ToList();

搜索返回以下文档

{ "_id" : 1, "make" : "Fender", "description" : "Classic guitars known for their versatility.", "establishedYear" : 1946, "in_stock" : true, "in_stock_location" : { "type" : "Point", "coordinates" : ["-73.93615", "40.69791"] }, "rating" : 9 }

有关 geoWithin 操作符的更多信息，请参阅 geoWithin Atlas 指南。

更多类似

使用MoreLikeThis()方法搜索与输入文档类似的文档。

以下示例搜索guitars集合中，描述字段值为"高质量"的对象类似的文档。

var searchDocument = new GuitarSearch()
{
    Description = "high quality",
};
var result = guitarsCollection.Aggregate()
    .Search(Builders<Guitar>.Search.MoreLikeThis(searchDocument))
    .ToList();

搜索返回以下文档

{ "_id" : 3, "make" : "PRS", "description" : "High-end guitars known for their quality.", "establishedYear" : 1985, "in_stock" : true, "rating" : 9 }
{ "_id" : 4, "make" : "Kiesel", "description" : "Quality guitars made only for custom orders.", "establishedYear" : 2015, "in_stock" : false, "rating" : null }

有关moreLikeThis运算符的更多信息，请参阅moreLikeThis Atlas指南。

接近

使用Near()方法搜索指定字段值接近给定值的文档。您可以在以下字段上执行搜索：

  • 数字字段

  • 日期字段

  • 地理点

以下示例搜索guitars集合中，评分字段值接近9的文档。文档根据其值接近9的远近顺序返回。

var result = guitarsCollection.Aggregate()
    .Search(Builders<Guitar>.Search.Near(g => g.Rating, 9, 1))
    .ToList();

搜索返回以下文档

{ "_id" : 1, "make" : "Fender", "description" : "...", "establishedYear" : 1946, "in_stock" : true, "rating" : 9 }
{ "_id" : 3, "make" : "PRS", "description" : "...", "establishedYear" : 1985, "in_stock" : true, "rating" : 9 }
{ "_id" : 2, "make" : "Gibson", "description" : "...", "establishedYear" : 1902, "in_stock" : true, "rating" : 8 }
{ "_id" : 5, "make" : "Ibanez", "description" : "...", "establishedYear" : 1957, "in_stock" : true, "rating" : 7 }

有关near运算符的更多信息，请参阅near Atlas指南。

短语

使用 Phrase() 方法来搜索包含指定字段和输入字符串的文档。

以下示例搜索 guitars 集合中，description 字段包含短语 "classic guitars" 的文档。

var result = guitarsCollection.Aggregate()
    .Search(Builders<Guitar>.Search.Phrase(g => g.Description, "classic guitars"))
    .ToList();

搜索返回以下文档

{ "_id" : 1, "make" : "Fender", "description" : "Classic guitars known for their versatility.", "establishedYear" : 1946, "in_stock" : true, "rating" : 9 }
{ "_id" : 2, "make" : "Gibson", "description" : "Classic guitars known for their rich, full tones.", "establishedYear" : 1902, "in_stock" : true, "rating" : 8 }

您还可以按照以下方式搜索匹配多个单独短语的集合中的文档

var result = guitarsCollection.Aggregate()
    .Search(Builders<Guitar>.Search.Phrase(g => g.Description, new List<string>() { "classic guitars", "quality guitars" }))
    .ToList();

此搜索返回以下文档

{ "_id" : 1, "make" : "Fender", "description" : "Classic guitars known for their versatility.", "establishedYear" : 1946, "in_stock" : true, "rating" : 9 }
{ "_id" : 4, "make" : "Kiesel", "description" : "Quality guitars made only for custom orders.", "establishedYear" : 2015, "in_stock" : false, "rating" : null }
{ "_id" : 2, "make" : "Gibson", "description" : "Classic guitars known for their rich, full tones.", "establishedYear" : 1902, "in_stock" : true, "rating" : 8 }

要了解有关 phrase 操作符的更多信息，请参阅phrase Atlas指南。

查询字符串

使用 QueryString() 方法通过以下运算符和分隔符进行字符串搜索

  • AND

  • OR

  • NOT

  • ()

以下示例搜索 guitars 集合中，description 字段的值满足以下每个条件的文档

  • 包含字符串 "classic" 或字符串 "quality" 或

  • 不包含字符串 "custom" 或

var result = guitarsCollection.Aggregate()
    .Search(Builders<Guitar>.Search.QueryString(g => g.Description, "(classic OR quality) AND NOT custom"))
    .ToList();

搜索返回以下文档

{ "_id" : 1, "make" : "Fender", "description" : "Classic guitars known for their versatility.", "establishedYear" : 1946, "in_stock" : true, "rating" : 9 }
{ "_id" : 3, "make" : "PRS", "description" : "High-end guitars known for their quality.", "establishedYear" : 1985, "in_stock" : true, "rating" : 9 }
{ "_id" : 2, "make" : "Gibson", "description" : "Classic guitars known for their rich, full tones.", "establishedYear" : 1902, "in_stock" : true, "rating" : 8 }

要了解有关 queryString 操作符的更多信息，请参阅queryString Atlas指南。

范围

使用 Range() 方法搜索指定字段的值在给定数字或日期范围内的文档。

以下示例搜索 guitars 集合中所有 establishedYear 值大于 1980 且小于 2020 的文档。

var result = guitarsCollection.Aggregate()
    .Search(Builders<Guitar>.Search.Range(g => g.EstablishedYear, SearchRangeBuilder.Gt(1980).Lt(2020)))
    .ToList();

搜索返回以下结果

{ "_id" : 3, "make" : "PRS", "description" : "High-end guitars known for their quality.", "establishedYear" : 1985, "in_stock" : true, "rating" : 9 }
{ "_id" : 4, "make" : "Kiesel", "description" : "Quality guitars made only for custom orders.", "establishedYear" : 2015, "in_stock" : false, "rating" : null }
{ "_id" : 6, "make" : "Strandberg", "description" : "Modern guitars known for their headless models.", "establishedYear" : 1982, "in_stock" : false, "rating" : null }

要了解有关 range 操作符的更多信息，请参阅range Atlas指南。

正则表达式

使用Regex()方法使用正则表达式搜索文档。

以下示例搜索guitars集合中，make字段的值恰好包含六个字母的文档。

var regex = "[A-Za-z]{6}";
var result = guitarsCollection.Aggregate()
    .Search(Builders<Guitar>.Search.Regex(g => g.Make, regex))
    .ToList();

搜索返回以下结果

{ "_id" : 1, "make" : "Fender", "description" : "Classic guitars known for their versatility.", "establishedYear" : 1946, "in_stock" : true, "rating" : 9 }
{ "_id" : 2, "make" : "Gibson", "description" : "Classic guitars known for their rich, full tones.", "establishedYear" : 1902, "in_stock" : true, "rating" : 8 }
{ "_id" : 4, "make" : "Kiesel", "description" : "Quality guitars made only for custom orders.", "establishedYear" : 2015, "in_stock" : false, "rating" : null }
{ "_id" : 5, "make" : "Ibanez", "description" : "Well-crafted guitars used by many professional guitarists.", "establishedYear" : 1957, "in_stock" : true, "rating" : 7 }

注意

默认情况下，regex运算符不能在分析字段上运行。您可以通过将allowAnalyzedField选项设置为true来允许它在分析字段上运行，如下所示

var result = guitarsCollection.Aggregate()
 .Search(Builders<Guitar>.Search.Regex(g => g.Make, regex, true))
 .ToList();

将allowAnalyzedField选项设置为true可能会导致意外的搜索结果。有关更多信息，请参阅regex行为。

有关regex运算符的更多信息，请参阅regex Atlas指南。

跨度

使用Span()方法在字段的区域中搜索文本搜索匹配项。您可以使用此方法查找彼此非常接近的字符串。

注意

span运算符比其他运算符更复杂，因为查询必须跟踪位置信息。

以下示例搜索guitars集合中，description字段的值包含“guitars”和“quality”字符串，这两个字符串彼此之间在一个单词之内。

var searchTerms = new[]
{
     Builders<Guitar>.SearchSpan.Term(g => g.Description, "guitars"),
     Builders<Guitar>.SearchSpan.Term(g => g.Description, "quality")
 };
var result = guitarsCollection.Aggregate()
    .Search(Builders<Guitar>.Search.Span(Builders<Guitar>.SearchSpan.Near(searchTerms, 1)))
    .ToList();

搜索返回以下文档

{ "_id" : 4, "make" : "Kiesel", "description" : "Quality guitars made only for custom orders.", "establishedYear" : 2015, "in_stock" : false, "rating" : null }

尽管文档中包含_id: 3的字符串“guitars”和“quality”，但它们被分隔在一个单词以上，因此搜索将此文档排除在结果之外。

有关span运算符的更多信息，请参阅span Atlas指南。

文本

使用Text()方法在文档中搜索给定的字符串或字符串数组。如果给定字符串中有多个术语，Atlas Search也会分别搜索字符串中的每个术语的匹配项。

以下示例搜索guitars集合中，其中description字段值包含字符串"used by professionals"的文档。

var result = guitarsCollection.Aggregate()
    .Search(Builders<Guitar>.Search.Text(g => g.Description, "used by professional"))
    .ToList();

搜索返回以下文档

{ "_id" : 5, "make" : "Ibanez", "description" : "Well-crafted guitars used by many professional guitarists.", "establishedYear" : 1957, "in_stock" : true, "rating" : 7 }

提示

如果您的搜索字符串包含多个术语，该方法还会分别搜索字符串中的每个术语的匹配项。

有关text运算符的更多信息，请参阅text Atlas指南。

通配符

使用Wildcard()方法通过搜索字符串中的特殊字符来搜索文档，这些特殊字符可以匹配任何字符。您可以在搜索中使用以下字符

字符
描述

?

匹配任何单个字符

*

匹配0个或多个字符

\

转义字符

以下示例搜索在make字段值包含字符串"Strand"后跟任意字符的文档。

var result = guitarsCollection.Aggregate()
    .Search(Builders<Guitar>.Search.Wildcard(g => g.Make, "Strand*"))
    .ToList();

搜索返回以下文档

{ "_id" : 6, "make" : "Strandberg", "description" : "Modern guitars known for their headless models.", "establishedYear" : 1982, "in_stock" : false, "rating" : null }

注意

默认情况下，wildcard运算符不能在分析字段上运行。您可以通过将allowAnalyzedField选项设置为true来允许它在不分析字段上运行，如下所示

var result =  guitarsCollection.Aggregate()
 .Search(Builders<Guitar>.Search.Wildcard(g => g.Make, "Strand*", true))
 .ToList();

将allowAnalyzedField选项设置为true可能会导致意外的搜索结果。有关更多信息，请参阅通配符行为。

有关wildcard运算符的更多信息，请参阅通配符 Atlas指南。

返回

使用构建器的操作

下一步

稳定API