Configuring Data Lake

Data Lake Configuration¶

Overview¶

The Atlas Data Lake configuration is in JSON format. It contains mappings between your data stores and Data Lake. Data Lake supports S3 buckets, Atlas clusters, and publicly accessible URL s as data stores. You must define mappings in your Data Lake to your S3 bucket, Atlas cluster, and HTTP data stores to run queries.

Example Configuration for Individual Data Stores¶

Click on the tab below to learn more about the Data Lake configuration for that data store.

Example

Consider a S3 bucket datacenter-alpha containing data collected from a datacenter:

|--metrics
  |--hardware

The /metrics/hardware path stores JSON files with metrics derived from the datacenter hardware, where each filename is the UNIX timestamp in milliseconds of the 24 hour period covered by that file:

/hardware/1564671291998.json

The following configuration:

Defines a data store on the datacenter-alpha S3 bucket in the us-east-1 AWS region. The data store is specifically restricted to only datafiles in the metrics folder path.
Maps files from the hardware folder to a MongoDB database datacenter-alpha-metrics and collection hardware. The configuration mapping includes parsing logic for capturing the timestamp implied in the filename.

{
  "stores" : [
    {
      "name" : "datacenter-alpha",
      "provider" : "s3",
      "region" : "us-east-1",
      "bucket" : "datacenter-alpha",
      "additionalStorageClasses" : [
        "STANDARD_IA"
      ],
      "prefix" : "/metrics",
      "delimiter" : "/"
    }
  ],
  "databases" : [
    {
      "name" : "datacenter-alpha-metrics",
      "collections" : [
        {
          "name" : "hardware",
          "dataSources" : [
            {
              "storeName" : "datacenter-alpha",
              "path" : "/hardware/{date date}"
            }
          ]
        }
      ]
    }
  ]
}

Atlas Data Lake parses the S3 bucket datacenter-alpha and processes all files under /metrics/hardware/. The collections uses the path parsing syntax to map the filename to the date field, which is an ISO-8601 date, in each document. If a matching date field does not exist in a document, it will be added.

Users connected to the Data Lake can use the MongoDB Query Language and supported aggregations to analyze data in the S3 bucket through the datacenter-metrics.hardware collection.

Tip

See Also:

Querying Data on S3

Example Configuration for Running Federated Queries¶

You can define mappings between your S3 , Atlas cluster, and HTTP data stores and Data Lake in the storage configuration to run federated queries.

Example

For the preceding sample S3 , Atlas cluster, and HTTP data stores, the Data Lake configuration for federated queries resembles the following:

{
  "stores" : [
    {
      "name" : "datacenter-alpha",
      "provider" : "s3",
      "region" : "us-east-1",
      "bucket" : "datacenter-alpha",
      "additionalStorageClasses" : [
        "STANDARD_IA"
      ],
      "prefix" : "/metrics",
      "delimiter" : "/"
    },
    {
      "name" : "atlasClusterStore",
      "provider" : "atlas",
      "clusterName" : "myDataCenter",
      "projectId" : "5e2211c17a3e5a48f5497de3"
    },
    {
      "name" : "httpStore",
      "provider" : "http",
      "allowInsecure" : false,
      "urls" [
        "https://www.datacenter-hardware.com/data.json",
        "https://www.datacenter-software.com/data.json"
      ],
      "defaultFormat" : ".json"
    }
  ],
  "databases" : [
    {
      "name" : "datacenter-metrics",
      "collections" : [
        {
          "name" : "inventory",
          "dataSources" : [
            {
              "storeName" : "datacenter-alpha",
              "path" : "/hardware/{date date}"
            },
            {
              "storeName" : "atlasClusterStore",
              "database" : "metrics",
              "collection" : "hardware"
            },
            {
              "storeName" : "httpStore",
              "allowInsecure" : false,
              "urls": [
                "https://www.datacenter-metrics.com/data.json"
              ],
              "defaultFormat" : ".json"
            }
          ]
        }
      ]
    }
  ]
}

Important

If the database in the storage configuration contains collections from S3 , Atlas, and HTTP data stores, the query results might contain data from all the data stores.

Tip

See Also:

Running Federated Queries

Configuration Format¶

The Data Lake configuration has the following format:

{
  "stores" : [
    {
      "name" : "<string>",
      "provider": "<string>",
      "region" : "<string>",
      "bucket" : "<string>",
      "additionalStorageClasses" : ["<string>"],
      "prefix" : "<string>",
      "includeTags": <boolean>,
      "delimiter": "<string>"
    }
  ],
    "databases" : [
      {
        "name" : "<string>",
        "collections" : [
          {
            "name" : "<string>",
            "dataSources" : [
              {
                "storeName" : "<string>",
                "path" : "<string>",
                "defaultFormat" : "<string>"
              }
            ]
          }
        ],
        "maxWildcardCollections" : <integer>,
        "views" : [
          {
            "name" : "<string>",
            "source" : "<string>",
            "pipeline" : "<string>"
          }
        ]
      }
    ]
  }

stores: The stores object defines each data store associated to the Data Lake. Data Lake can only access data stores defined in the stores object.
databases: The databases object defines the mapping between each data store defined in stores and MongoDB collections in the databases.

`stores`¶

"stores" : [
  {
    "name" : "<string>",
    "provider" : "<string>",
    "region" : "<string>",
    "bucket" : "<string>",
    "additionalStorageClasses" : ["<string>"],
    "prefix" : "<string>",
    "delimiter" : "<string>",
    "includeTags": <boolean>
  }
]

stores¶: The stores object defines an array of data stores associated with a Data Lake. The data store captures files in an S3 bucket, documents in Atlas cluster, or files stored at publicly accessible URL s. An Atlas Data Lake can only access data stores defined in the stores object.

stores.[n].name¶: Name of the data store. The databases.[n].collections.[n].dataSources.[n].storeName field references this value as part of mapping configuration.

stores.[n].provider¶

Defines where the data is stored. Value can be one of the following:

s3 for an AWS S3 bucket.
atlas for a collection in an Atlas cluster.
http for data in files hosted at publicly accessible URL s.

stores.[n].region¶: Name of the AWS region in which the S3 bucket is hosted. For a list of valid region names, see Amazon Web Services (AWS).

stores.[n].bucket¶: Name of the AWS S3 bucket. Must exactly match the name of an S3 bucket which Data Lake can access given the configured AWS IAM credentials.

stores.[n].additionalStorageClasses¶

Optional. Array of AWS S3 storage classes. Atlas Data Lake will include the files in these storage classes in the query results. Valid values are:

INTELLIGENT_TIERING to include files in the Intelligent Tiering storage class
STANDARD_IA to include files in the Standard-Infrequent Access storage class

Note

Files in the Standard storage class are supported by default.

stores.[n].prefix¶

Optional. Prefix Data Lake applies when searching for files in the S3 bucket.

For example, consider an S3 bucket metrics with the following structure:

metrics
  |--hardware
  |--software
     |--computed

The data store prepends the value of prefix to the databases.[n].collections.[n].dataSources.[n].path to create the full path for files to ingest. Setting the prefix to /software restricts any databases objects using the data store to only subpaths /software.

If omitted, Data Lake searches all files from the root of the S3 bucket.

stores.[n].delimiter¶

Optional. The delimiter that separates databases.[n].collections.[n].dataSources.[n].path segments in the data store. Data Lake uses the delimiter to efficiently traverse S3 buckets with a hierarchical directory structure. You can specify any character supported by the S3 object keys as the delimiter. For example, you can specify an underscore (_) or a plus sign (+) or multiple characters such as double underscores (__) as the delimiter.

If omitted, defaults to "/".

stores.[n].includeTags¶

Optional. Determines whether or not to use S3 tags on the files in the given path as additional partition attributes. Valid values are true and false.

If omitted, defaults to false.

If set to true, Data Lake does the following:

Adds the S3 tags as additional partition attributes.
Adds new top level BSON elements that associates each tag to each document for the tagged files.

Warning

If set to true, Data Lake processes the files for additional partition attributes by making extra calls to S3 to get the tags. This behavior might impact performance.

`databases`¶

"databases" : [
  {
    "name" : "<string>",
    "collections" : [
      {
        "name" : "<string>",
        "dataSources" : [
          {
            "storeName" : "<string>",
            "defaultFormat" : "<string>",
            "path" : "<string>"
          }
        ]
      }
    ],
    "maxWildcardCollections" : <integer>,
    "views" : [
      {
        "name" : "<string>",
        "source" : "<string>",
        "pipeline" : "<string>"
      }
    ]
  }
]

databases¶: Array of objects where each object represents a database, its collections, and, optionally, any views on the collections. Each database can have multiple collections and views objects.

databases.[n].name¶: Name of the database to which Data Lake maps the data contained in the data store.

databases.[n].collections¶: Array of objects where each object represents a collection and data sources that map to a stores data store.

databases.[n].collections.name¶: Name of the collection to which Data Lake maps the data contained in each databases.[n].collections.[n].dataSources.[n].storeName. Each object in the array represents the mapping between the collection and an object in the stores array.
You can generate collection names dynamically from file paths by specifying * for the collection name and the collectionName() function in the databases.[n].collections.[n].dataSources.[n].path field. See Generate Dynamic Collection Names from File Path for examples.

databases.[n].collections.[n].dataSources¶: Array of objects where each object represents a stores data store to map with the collection.

databases.[n].collections.[n].dataSources.[n].storeName¶: Name of a data store to map to the <collection>. Must match the stores.[n].name of an object in the stores array.

databases.[n].collections.[n].dataSources.[n].path¶

Controls how Atlas Data Lake searches for and parses files in the databases.[n].collections.[n].dataSources.[n].storeName before mapping them to the <collection>. Data Lake prepends the stores.[n].prefix to the path to build the full path to search within. Specify / to capture all files and folders from the prefix path.

For example, consider an S3 bucket metrics with the following structure:

metrics
|--hardware
|--software
   |--computed

A path of / directs Data Lake to search all files and folders in the metrics bucket.

A path of /hardware directs Data Lake to search only that path for files to ingest.

If the stores.[n].prefix is software, Data Lake searches for files only in the path /software/computed.

Appending the * wildcard character to the path directs Data Lake to include all files and folders from that point in the path. For example, /software/computed* would match files like /software/computed-detailed, /software/computedArchive, and /software/computed/errors.

databases.[n].collections.[n].dataSources.[n].path supports additional syntax for parsing filenames, including:

Generating document fields from filenames.
Using regular expressions to control field generation.
Setting boundaries for bucketing filenames by timestamp.

See Path Syntax for more information.

When specifying the databases.[n].collections.[n].dataSources.[n].path:

Specify the data type for the partition attribute.
Ensure that the partition attribute type matches the data type to parse.
Use the delimiter specified in stores.[n].delimiter.

databases.[n].collections.[n].dataSources.[n].defaultFormat¶

Optional. Default format that Data Lake assumes if it encounters a file without an extension while searching the databases.[n].collections.[n].dataSources.[n].storeName.

The following values are valid for the defaultFormat field:

.json, .json.gz, .bson, .bson.gz, .avro, .avro.gz, .orc, .tsv, .tsv.gz, .csv, .csv.gz, .parquet

Note

If your file format is CSV or TSV, you must include a header row in your data. See Comma-Separated and Tab-Separated Value Data Files for more information.

If omitted, Data Lake attempts to detect the file type by processing a few bytes of the file.

Tip

See Also:

Supported Data Formats

databases.[n].maxWildcardCollections¶: Optional. Maximum number of wildcard * collections in the database. Each wildcard collection can have only one data source. Value can be between 1 and 1000, inclusive. If omitted, defaults to 100.

databases.[n].views¶: Array of objects where each object represents an aggregation pipeline on a collection. To learn more about views, see Views.

databases.[n].views.[n].name¶: Name of the view.

databases.[n].views.[n].source¶: Name of the source collection for the view.

databases.[n].views.[n].pipeline¶: Aggregation pipeline stage(s) to apply to the databases.[n].views.[n].source collection.

Tip

See Also:

Give Feedback

← Configuring Data Lake Path Syntax →

Data Lake Configuration¶

Overview¶

Example Configuration for Individual Data Stores¶

Example Configuration for Running Federated Queries¶

Configuration Format¶

stores¶

databases¶

`stores`¶

`databases`¶