本文属于机器翻译版本。若本译文内容与英语原文存在差异，则一律以英文原文为准。

# \$1\$1KEEP
<a name="KEEP"></a>

`$$KEEP`系统变量与聚合管道中的`$redact`阶段一起使用，以保持当前文档或字段不变，并将其包含在输出中。

**参数**

无

## 示例（MongoDB 外壳）
<a name="KEEP-examples"></a>

以下示例演示了在 Amazon DocumentDB 聚合管道`$$KEEP`中的用法。只有当访问权限等于 “公开” 时，才会保留文档，否则它们将被删除。

**创建示例文档**

```
db.articles.insertMany([
  { title: "Article A", access: "public", content: "Visible content" },
  { title: "Article B", access: "private", content: "Hidden content" }
]);
```

**查询示例**

```
db.articles.aggregate([
  {
    $redact: {
      $cond: [
        { $eq: ["$access", "public"] },
        "$$KEEP",
        "$$PRUNE"
      ]
    }
  }
]);
```

**输出**

```
[
  {
    "_id" : ObjectId("..."),
    "title" : "Article A",
    "access" : "public",
    "content" : "Visible content"
  }
]
```

## 代码示例
<a name="KEEP-code"></a>

要查看使用该`$$KEEP`命令的代码示例，请选择要使用的语言的选项卡：

------
#### [ Node.js ]

```
const { MongoClient } = require('mongodb');

async function run() {
  const client = new MongoClient(
    'mongodb://<username>:<password>@<cluster-endpoint>:27017/?tls=true&tlsCAFile=global-bundle.pem&replicaSet=rs0'
  );

  try {
    await client.connect();
    const db = client.db('test');
    const articles = db.collection('articles');

    const pipeline = [
      {
        $redact: {
          $cond: [
            { $eq: ["$access", "public"] },
            "$$KEEP",
            "$$PRUNE"
          ]
        }
      }
    ];

    const results = await articles.aggregate(pipeline).toArray();
    console.log(results);
  } finally {
    await client.close();
  }
}

run().catch(console.error);
```

------
#### [ Python ]

```
from pymongo import MongoClient

client = MongoClient(
    "mongodb://<username>:<password>@<cluster-endpoint>:27017/?tls=true&tlsCAFile=global-bundle.pem&replicaSet=rs0"
)

db = client.test
articles = db.articles

pipeline = [
    {
        "$redact": {
            "$cond": [
                {"$eq": ["$access", "public"]},
                "$$KEEP",
                "$$PRUNE"
            ]
        }
    }
]

results = list(articles.aggregate(pipeline))
print(results)

client.close()
```

------