Docs Menu
Docs Home
/
Database Manual
/
Aggregation Operations
/
Reference
/
Operators

$meta

Definition

$meta

Returns the metadata associated with a document, e.g. "textScore" when performing text search.

A $meta expression has the following syntax:

{ $meta: <metaDataKeyword> }

The $meta expression can specify the following values as the <metaDataKeyword>:

Keyword
Description

"textScore"

Returns the score associated with the corresponding $text query for each matching document. The text score signifies how well the document matched the search term or terms.

{ $meta: "textScore" } must be used in conjunction with a $text query.

In earlier versions, if not used in conjunction with a $text query, returns a score of null.

$text provides text query capabilities for self-managed (non-Atlas) deployments. For data hosted on MongoDB Atlas, MongoDB offers an improved full-text query solution, Atlas Search.

"indexKey"

Returns an index key for the document if a non-text index is used. The { $meta: "indexKey" } expression is for debugging purposes only, and not for application logic, and is preferred over cursor.returnKey().

MongoDB Atlas Search provides additional $meta keywords, such as:

Refer to the Atlas Search documentation for details.

Important

The following $meta keywords are not supported in Stable API V1:

  • "textScore"

  • "indexKey"

  • "searchScore"

  • "searchHighlights" (Available starting in MongoDB 7.0.5 and 6.0.13)

  • "searchSequenceToken"

  • "searchScoreDetails"

Behavior

Text Score Metadata $meta: "textScore"

Requires $text Search

  • The { $meta: "textScore" } expression must be used in conjunction with $text. For example:

    • In aggregation, you must specify a $match stage with a $text query in the pipeline to use the { $meta: "textScore" } expression in later stage(s). If you do not specify the $text query in the $match stage, the operation fails.

    • In find, you must specify the $text operator in the query predicate to use { $meta: "textScore" }. If you do not specify the $text operator in the query predicate, the operation fails.

    Note

    $text provides text query capabilities for self-managed (non-Atlas) deployments. For data hosted on MongoDB Atlas, MongoDB offers an improved full-text query solution, Atlas Search.

Availability

  • In aggregation, the { $meta: "textScore" } expression can be included in various stages that accept aggregation expressions, such as $project, $group $sort, etc.

  • In find, the { $meta: "textScore" } expression can be included in projection and in sort().

Usage In Projection

  • The { $meta: "textScore" } expression can be a part of the projection document to include the text score metadata.

  • The $meta expression can be present in either an inclusion or an exclusion projection.

  • If you set the expression to a field name that already exists in the document, the projected metadata value overwrites the existing value.

Filter on Text Score

  • In aggregation, following a stage that outputs a field with the text score value, you can specify a query condition or operate on the field in subsequent stages. For example, see $text in the Aggregation Pipeline on Self-Managed Deployments.

  • In find, you cannot specify a query condition on the text score. Use aggregation instead.

Usage In Sort

  • The { $meta: "textScore" } expression can be used as a part of a sort operation to sort by the text score metadata; i.e.,

  • The "textScore" metadata sorts in descending order.

  • To use in a sort operation, set the { $meta: "textScore" } expression to an arbitrary field name. The field name is disregarded by the query system.

Sort without Projection

  • In aggregation, you can sort the resulting documents by { $meta: "textScore" } without also having to project the textScore.

  • In find, you can sort the resulting documents by { $meta: "textScore" } without also having to project the textScore.

Sort with Projection

  • In aggregation, if you include the { $meta: "textScore" } expression in both the projection and sort, the projection and sort can have different field names for the expression. The field name in the sort is disregarded by the query system.

  • In find, if you include the { $meta: "textScore" } expression in both the projection and sort, the projection and sort can have different field names for the expression. The field name in the sort is disregarded by the query system.

Index Key Metadata $meta: "indexKey" (Aggregation and Find)

Usage

  • The { $meta: "indexKey" } expression is for debugging purposes only and not for application logic.

  • The { $meta: "indexKey" } expression is preferred over cursor.returnKey().

Availability

  • In aggregation, the { $meta: "indexKey" } expression can be included in various stages that accept aggregation expressions, such as $project, $group $sortByCount, etc., but not $sort. However, with an aggregation pipeline, you can first project the { $meta: "indexKey" } expression (such as in a $project, $addFields, etc. ) and then, sort by that field in a subsequent $sort stage.

  • In find, the { $meta: "indexKey" } expression is only available as part of the projection document.

Return Value

  • The value returned depends on how the database decides to represent values in an index and may change across versions. The represented value may not be the actual value for the field.

  • The value returned depends on the execution plan chosen by the system. For example, if there are two possible indexes which can be used to answer the query, then the value of the "indexKey" metadata depends on which index is selected.

  • If an index is not used, the { $meta: "indexKey" } expression does not return a value and the field is not included as part of the output.

Examples

$meta: "textScore"

Create an articles collection with the following documents:

db.articles.insertMany([
   { "_id" : 1, "title" : "cakes and ale" },
   { "_id" : 2, "title" : "more cakes" },
   { "_id" : 3, "title" : "bread" },
   { "_id" : 4, "title" : "some cakes" },
   { "_id" : 5, "title" : "two cakes to go" },
   { "_id" : 6, "title" : "pie" }
])

Create a text index on the title field:

db.articles.createIndex( { title: "text"} )

The following aggregation operation performs a text search and uses the $meta operator to group by the text search score:

db.articles.aggregate(
   [
     { $match: { $text: { $search: "cake" } } },
     { $group: { _id: { $meta: "textScore" }, count: { $sum: 1 } } }
   ]
)

The operation returns the following results:

{ "_id" : 0.75, "count" : 1 }
{ "_id" : 0.6666666666666666, "count" : 1 }
{ "_id" : 1, "count" : 2 }

For more examples, see $text in the Aggregation Pipeline on Self-Managed Deployments.

The following query performs a text search for the term cake and uses the $meta operator in the projection document to include the score assigned to each matching document:

db.articles.find(
   { $text: { $search: "cake" } },
   { score: { $meta: "textScore" } }
)

The operation returns the following documents with the text score:

{ "_id" : 4, "title" : "some cakes", "score" : 1 }
{ "_id" : 1, "title" : "cakes and ale", "score" : 0.75 }
{ "_id" : 5, "title" : "two cakes to go", "score" : 0.6666666666666666 }
{ "_id" : 2, "title" : "more cakes", "score" : 1 }

For additional examples of "textScore" projections and sorts, see Relevance Score Examples.

$meta: "indexKey"

Note

The { $meta: "indexKey" } expression is for debugging purposes only and not for application logic. MongoDB returns the value associated with the index chosen by the query system. The system can choose a different index upon subsequent execution.

For the selected index, the value returned depends on how the database decides to represent values in an index and may change across versions. The represented value may not be the actual value for the field.

Create an orders collection with the following documents:

db.orders.insertMany([
   { "item" : "abc", "price" : NumberDecimal("12"), "quantity" : 2, "type": "apparel" },
   { "item" : "jkl", "price" : NumberDecimal("20"), "quantity" : 1, "type": "electronics" },
   { "item" : "abc", "price" : NumberDecimal("10"), "quantity" : 5, "type": "apparel" }
])

Create the following compound index on the type and item fields:

db.orders.createIndex( { type: 1, item: 1 } )

The following aggregation operation finds all documents with type equal to apparel and uses the $meta operator to include the index key value for the matching document if an index was used:

db.orders.aggregate(
   [
     { $match: { type: "apparel" } },
     { $addFields: { idxKey: { $meta: "indexKey" } } }
   ]
)

The following operation finds all documents with type equal to apparel and uses the $meta operator to include the index key value for the matching document if an index was used:

db.orders.find( { type: "apparel" }, { idxKey: { $meta: "indexKey" } } )

The operation returns the matching documents with their corresponding index key:

{
   "_id" : ObjectId("5e98a33ceaf5e9dcf2b8dcde"),
   "item" : "abc",
   "price" : NumberDecimal("12"),
   "quantity" : 2,
   "type" : "apparel",
   "idxKey" : { "type" : "apparel", "item" : "abc" }
}
{
   "_id" : ObjectId("5e98a33ceaf5e9dcf2b8dce0"),
   "item" : "abc",
   "price" : NumberDecimal("10"),
   "quantity" : 5,
   "type" : "apparel",
   "idxKey" : { "type" : "apparel", "item" : "abc" }
}

If no index is used, the { $meta: "indexKey" } does not return anything.

For example, the following operation does not use an index since no index exists on the price field to support the match condition:

db.orders.aggregate(
   [
     { $match: { price: { $gte: NumberDecimal("10") } } },
     { $addFields: { idxKey: { $meta: "indexKey" } } }
   ]
)

For example, the following operation does not use an index since no index exists on the price field to support the match condition:

db.orders.find(
   { price: { $gte: NumberDecimal("10") } },
   { idxKey: { $meta: "indexKey" } }
)

The operation returns the matching documents without the idxKey field:

{
   "_id" : ObjectId("5e98a33ceaf5e9dcf2b8dcde"),
   "item" : "abc",
   "price" : NumberDecimal("12"),
   "quantity" : 2,
   "type" : "apparel"
}
{
   "_id" : ObjectId("5e98a33ceaf5e9dcf2b8dcdf"),
   "item" : "jkl",
   "price" : NumberDecimal("20"),
   "quantity" : 1,
   "type" : "electronics"
}
{
   "_id" : ObjectId("5e98a33ceaf5e9dcf2b8dce0"),
   "item" : "abc",
   "price" : NumberDecimal("10"),
   "quantity" : 5,
   "type" : "apparel"
}

The C# examples on this page use the sample_mflix database from the Atlas sample datasets. To learn how to create a free MongoDB Atlas cluster and load the sample datasets, see Get Started in the MongoDB .NET/C# Driver documentation.

The following Movie class models the documents in the sample_mflix.movies collection:

public class Movie
{
    public ObjectId Id { get; set; }
    public string Plot { get; set; }
    
    public List<string> Genres { get; set; }
    
    public int Runtime { get; set; }
    
    public List<string> Cast { get; set; }
    
    public string Title { get; set; }
    [BsonElement("lastupdated")]
    public DateTime LastUpdated { get; set; }
    
    public string Rated { get; set; }
    public int Year { get; set; }
    
    public ImdbData Imdb { get; set; }
    public string Type { get; set; }
    public int Index { get; set; }
    
    public string[] Comments { get; set; }
    public List<BsonDocument> Highlights { get; set; }
    
    public float Score { get; set; }
    [BsonElement("scoreDetails")]
    public SearchScoreDetails ScoreDetails { get; set; }
    
    [BsonElement("searchScoreDetails")]
    public SearchScoreDetails SearchScoreDetails { get; set; }
    
    [BsonElement("paginationToken")]
    public string PaginationToken { get; set; }
    
    [BsonElement("plot_embedding")]
    public float[] PlotEmbedding { get; set; }
}

Note

ConventionPack for Pascal Case

The properties in the preceding class are named in Pascal case, but the field names in the MongoDB collection use camel case. To account for this difference, you can use the following code to register a ConventionPack when your application starts:

var camelCaseConvention = new ConventionPack { new CamelCaseElementNameConvention() };
ConventionRegistry.Register("CamelCase", camelCaseConvention, type => true);

To include the text search score by using the .NET/C# driver, call the MetaTextScore() method on the projection builder.

First, run the following code to create a text index on the Title field:

var indexModel = new CreateIndexModel<Movie>(
    Builders<Movie>.IndexKeys.Text(m => m.Title));
movieCollection.Indexes.CreateOne(indexModel);

Then, add the following code to your application. The code performs a text search for the term "future" and uses the MetaTextScore() method to include the score assigned to each matching document. The code then sorts the results in descending order by their text score and returns the document with the highest score.

var filter = Builders<Movie>.Filter.Text("future");
var projection = Builders<Movie>.Projection
    .Include(m => m.Title)
    .Include(m => m.Plot)
    .MetaTextScore(m => m.Score);
var result = movieCollection.Find(filter)
    .Sort(Builders<Movie>.Sort.MetaTextScore("score"))
    .Project(projection)
    .FirstOrDefault();

The preceding operation returns the following document:

{
  "_id": {
    "$oid": "573a1398f29313caabceb500"
  },
  "plot": "After visiting 2015, Marty McFly must repeat his visit to 1955 to
   prevent disastrous changes to 1985... without interfering with his first trip.",
  "title": "Back to the Future Part II",
  "score": 1.59375
}

Back

$mergeObjects

Next

$min

On this page