Neptune Loader Command

• source   –   An Amazon S3 URI.

  The SOURCE parameter accepts an Amazon S3 URI that identifies a single file, multiple files, a folder, or multiple folders. Neptune loads every data file in any folder that is specified.

  The URI can be in any of the following formats.

  • s3://bucket_name/object-key-name

  • https://s3.amazonaws.com/bucket_name/object-key-name

  • https://s3.us-east-1.amazonaws.com/bucket_name/object-key-name

  The object-key-name element of the URI is equivalent to the prefix parameter in an Amazon S3 ListObjects API call. It identifies all the objects in the specified Amazon S3 bucket whose names begin with that prefix. That can be a single file or folder, or multiple files and/or folders.

  The specified folder or folders can contain multiple vertex files and multiple edge files.

  For example, if you had the following folder structure and files in an Amazon S3 bucket named bucket-name:

  s3://bucket-name/a/bc
  s3://bucket-name/ab/c
  s3://bucket-name/ade
  s3://bucket-name/bcd

  If the source parameter is specified as s3://bucket-name/a, the first three files will be loaded.

  s3://bucket-name/a/bc
  s3://bucket-name/ab/c
  s3://bucket-name/ade

• format   –   The format of the data. For more information about data formats for the Neptune Loader command, see Using the Amazon Neptune Bulk Loader to Ingest Data.

  Allowed values

  • csv for the Gremlin CSV data format.

  • opencypher for the openCypher CSV data format.

  • ntriples for the N-Triples RDF data format.

  • nquads for the N-Quads RDF data format.

  • rdfxml for the RDF\XML RDF data format.

  • turtle for the Turtle RDF data format.

• iamRoleArn   –   The Amazon Resource Name (ARN) for an IAM role to be assumed by the Neptune DB instance for access to the S3 bucket. For information about creating a role that has access to Amazon S3 and then associating it with a Neptune cluster, see Prerequisites: IAM Role and Amazon S3 Access.

  Starting with engine release 1.2.1.0.R3, you can also chain multiple IAM roles if the Neptune DB instance and the Amazon S3 bucket are located in different Amazon Accounts. In this case, iamRoleArn contains a comma-separated list of role ARNs, as described in Chaining IAM roles in Amazon Neptune. For example:

  curl -X POST https://localhost:8182/loader \
    -H 'Content-Type: application/json' \
    -d '{
          "source" : "s3://(the target bucket name)/(the target date file name)",
          "iamRoleArn" : "arn:aws:iam::(Account A ID):role/(RoleA),arn:aws:iam::(Account B ID):role/(RoleB),arn:aws:iam::(Account C ID):role/(RoleC)",
          "format" : "csv",
          "region" : "us-east-1"
        }'

• region   –   The region parameter must match the Amazon Region of the cluster and the S3 bucket.

  Amazon Neptune is available in the following Regions:

  • US East (N. Virginia):   us-east-1

  • US East (Ohio):   us-east-2

  • US West (N. California):   us-west-1

  • US West (Oregon):   us-west-2

  • Canada (Central):   ca-central-1

  • South America (São Paulo):   sa-east-1

  • Europe (Stockholm):   eu-north-1

  • Europe (Ireland):   eu-west-1

  • Europe (London):   eu-west-2

  • Europe (Paris):   eu-west-3

  • Europe (Frankfurt):   eu-central-1

  • Middle East (Bahrain):   me-south-1

  • Middle East (UAE):   me-central-1

  • Israel (Tel Aviv):   il-central-1

  • Africa (Cape Town):   af-south-1

  • Asia Pacific (Hong Kong):   ap-east-1

  • Asia Pacific (Tokyo):   ap-northeast-1

  • Asia Pacific (Seoul):   ap-northeast-2

  • Asia Pacific (Osaka):   ap-northeast-3

  • Asia Pacific (Singapore):   ap-southeast-1

  • Asia Pacific (Sydney):   ap-southeast-2

  • Asia Pacific (Mumbai):   ap-south-1

  • China (Beijing):   cn-north-1

  • China (Ningxia):   cn-northwest-1

  • Amazon GovCloud (US-West):   us-gov-west-1

  • Amazon GovCloud (US-East):   us-gov-east-1

• mode   –   The load job mode.

  Allowed values: RESUME, NEW, AUTO.

  Default value: AUTO

  • RESUME   –   In RESUME mode, the loader looks for a previous load from this source, and if it finds one, resumes that load job. If no previous load job is found, the loader stops.

    The loader avoids reloading files that were successfully loaded in a previous job. It only tries to process failed files. If you dropped previously loaded data from your Neptune cluster, that data is not reloaded in this mode. If a previous load job loaded all files from the same source successfully, nothing is reloaded, and the loader returns success.

  • NEW   –   In NEW mode, the creates a new load request regardless of any previous loads. You can use this mode to reload all the data from a source after dropping previously loaded data from your Neptune cluster, or to load new data available at the same source.

  • AUTO   –   In AUTO mode, the loader looks for a previous load job from the same source, and if it finds one, resumes that job, just as in RESUME mode.

    If the loader doesn't find a previous load job from the same source, it loads all data from the source, just as in NEW mode.

• failOnError   –   A flag to toggle a complete stop on an error.

  Allowed values: "TRUE", "FALSE".

  Default value: "TRUE".

  When this parameter is set to "FALSE", the loader tries to load all the data in the location specified, skipping any entries with errors.

  When this parameter is set to "TRUE", the loader stops as soon as it encounters an error. Data loaded up to that point persists.

• parallelism   –   This is an optional parameter that can be set to reduce the number of threads used by the bulk load process.

  Allowed values:

  • LOW –   The number of threads used is the number of available vCPUs divided by 8.

  • MEDIUM –   The number of threads used is the number of available vCPUs divided by 2.

  • HIGH –   The number of threads used is the same as the number of available vCPUs.

  • OVERSUBSCRIBE –   The number of threads used is the number of available vCPUs multiplied by 2. If this value is used, the bulk loader takes up all available resources.

    This does not mean, however, that the OVERSUBSCRIBE setting results in 100% CPU utilization. Because the load operation is I/O bound, the highest CPU utilization to expect is in the 60% to 70% range.

  Default value: HIGH

  The parallelism setting can sometimes result in a deadlock between threads when loading openCypher data. When this happens, Neptune returns the LOAD_DATA_DEADLOCK error. You can generally fix the issue by setting parallelism to a lower setting and retrying the load command.

• parserConfiguration   –   An optional object with additional parser configuration values. Each of the child parameters is also optional:

  Name	Example Value	Description
  namedGraphUri	http://aws.amazon.com/neptune/vocab/v01/DefaultNamedGraph	The default graph for all RDF formats when no graph is specified (for non-quads formats and NQUAD entries with no graph). The default is http://aws.amazon.com/neptune/vocab/v01/DefaultNamedGraph
  baseUri	http://aws.amazon.com/neptune/default	The base URI for RDF/XML and Turtle formats. The default is http://aws.amazon.com/neptune/default.
  allowEmptyStrings	true	Gremlin users need to be able to pass empty string values("") as node and edge properties when loading CSV data. If allowEmptyStrings is set to false (the default), such empty strings are treated as nulls and are not loaded. If allowEmptyStrings is set to true, the loader treats empty strings as valid property values and loads them accordingly.

  For more information, see SPARQL Default Graph and Named Graphs.

• updateSingleCardinalityProperties   –   This is an optional parameter that controls how the bulk loader treats a new value for single-cardinality vertex or edge properties. This is not supported for loading openCypher data (see Loading openCypher data).

  Allowed values: "TRUE", "FALSE".

  Default value: "FALSE".

  By default, or when updateSingleCardinalityProperties is explicitly set to "FALSE", the loader treats a new value as an error, because it violates single cardinality.

  When updateSingleCardinalityProperties is set to "TRUE", on the other hand, the bulk loader replaces the existing value with the new one. If multiple edge or single-cardinality vertex property values are provided in the source file(s) being loaded, the final value at the end of the bulk load could be any one of those new values. The loader only guarantees that the existing value has been replaced by one of the new ones.

• queueRequest   –   This is an optional flag parameter that indicates whether the load request can be queued up or not.

  You don't have to wait for one load job to complete before issuing the next one, because Neptune can queue up as many as 64 jobs at a time, provided that their queueRequest parameters are all set to "TRUE". The queue order of the jobs will be first-in-first-out (FIFO).

  If the queueRequest parameter is omitted or set to "FALSE", the load request will fail if another load job is already running.

  Allowed values: "TRUE", "FALSE".

  Default value: "FALSE".

• dependencies   –   This is an optional parameter that can make a queued load request contingent on the successful completion of one or more previous jobs in the queue.

  Neptune can queue up as many as 64 load requests at a time, if their queueRequest parameters are set to "TRUE". The dependencies parameter lets you make execution of such a queued request dependent on the successful completion of one or more specified previous requests in the queue.

  For example, if load Job-A and Job-B are independent of each other, but load Job-C needs Job-A and Job-B to be finished before it begins, proceed as follows:

  1. Submit load-job-A and load-job-B one after another in any order, and save their load-ids.

  2. Submit load-job-C with the load-ids of the two jobs in its dependencies field:

    "dependencies" : ["job_A_load_id", "job_B_load_id"]

  Because of the dependencies parameter, the bulk loader will not start Job-C until Job-A and Job-B have completed successfully. If either one of them fails, Job-C will not be executed, and its status will be set to LOAD_FAILED_BECAUSE_DEPENDENCY_NOT_SATISFIED.

  You can set up multiple levels of dependency in this way, so that the failure of one job will cause all requests that are directly or indirectly dependent on it to be cancelled.

• userProvidedEdgeIds   –   This parameter is required only when loading openCypher data that contains relationship IDs. It must be included and set to True when openCypher relationship IDs are explicitly provided in the load data (recommended).

  When userProvidedEdgeIds is absent or set to True, an :ID column must be present in every relationship file in the load.

  When userProvidedEdgeIds is present and set to False, relationship files in the load must not contain an :ID column. Instead, the Neptune loader automatically generates an ID for each relationship.

  It's useful to provide relationship IDs explicitly so that the loader can resume loading after error in the CSV data have been fixed, without having to reload any relationships that have already been loaded. If relationship IDs have not been explicitly assigned, the loader cannot resume a failed load if any relationship file has had to be corrected, and must instead reload all the relationships.

• accessKey   –   [deprecated] An access key ID of an IAM role with access to the S3 bucket and data files.

  The iamRoleArn parameter is recommended instead. For information about creating a role that has access to Amazon S3 and then associating it with a Neptune cluster, see Prerequisites: IAM Role and Amazon S3 Access.

  For more information, see Access keys (access key ID and secret access key).

• secretKey   –   [deprecated] The iamRoleArn parameter is recommended instead. For information about creating a role that has access to Amazon S3 and then associating it with a Neptune cluster, see Prerequisites: IAM Role and Amazon S3 Access.

  For more information, see Access keys (access key ID and secret access key).

• When loading openCypher data in CSV format, the format parameter must be set to opencypher.

• The updateSingleCardinalityProperties parameter is not supported for openCypher loads because all openCypher properties have single cardinality. The openCypher load format does not support arrays, and if an ID value appears more than once, it is treated as a duplicate or an insertion error (see below).

• The Neptune loader handles duplicates that it encounters in openCypher data as follows:

  • If the loader encounters multiple rows with the same node ID, they are merged using the following rule:

    • All the labels in the rows are added to the node.

    • For each property, only one of the property values is loaded. The selection of the one to load is non-deterministic.

  • If the loader encounters multiple rows with the same relationship ID, only one of them is loaded. The selection of the one to load is non-deterministric.

  • The loader never updates property values of an existing node or relationship in the database if it encounters load data having the ID of the existing node or relationship. However, it does load node labels and properties that are not present in the existing node or relationship.

• Although you don't have to assign IDs to relationships, it is usually a good idea (see the userProvidedEdgeIds parameter above). Without explicit relationship IDs, the loader must reload all relationships in case of an error in a relationship file, rather than resuming the load from where it failed.

  Also, if the load data doesn't contain explicit relationship IDs, the loader has no way of detecting duplicate relationships.

Error Categories

• Error 400   –   Syntax errors return an HTTP 400 bad request error. The message describes the error.

• Error 500   –   A valid request that cannot be processed returns an HTTP 500 internal server error. The message describes the error.

Loader Error Messages

• Couldn't find the Amazon credential for iam_role_arn  (HTTP 400)

  The credentials were not found. Verify the supplied credentials against the IAM console or Amazon CLI output. Make sure that you have added the IAM role specified in iamRoleArn to the cluster.

• S3 bucket not found for source  (HTTP 400)

  The S3 bucket does not exist. Check the name of the bucket.

• The source source-uri does not exist/not reachable  (HTTP 400)

  No matching files were found in the S3 bucket.

• Unable to connect to S3 endpoint. Provided source = source-uri and region = aws-region  (HTTP 500)

  Unable to connect to Amazon S3. Region must match the cluster Region. Ensure that you have a VPC endpoint. For information about creating a VPC endpoint, see Creating an Amazon S3 VPC Endpoint.

• Bucket is not in provided Region (aws-region)  (HTTP 400)

  The bucket must be in the same Amazon Region as your Neptune DB instance.

• Unable to perform S3 list operation  (HTTP 400)

  The IAM user or role provided does not have List permissions on the bucket or the folder. Check the policy or the access control list (ACL) on the bucket.

• Start new load operation not permitted on a read replica instance  (HTTP 405)

  Loading is a write operation. Retry load on the read/write cluster endpoint.

• Failed to start load because of unknown error from S3  (HTTP 500)

  Amazon S3 returned an unknown error. Contact Amazon Web Services Support.

• Invalid S3 access key  (HTTP 400)

  Access key is invalid. Check the provided credentials.

• Invalid S3 secret key  (HTTP 400)

  Secret key is invalid. Check the provided credentials.

• Max concurrent load limit breached  (HTTP 400)

  If a load request is submitted without "queueRequest" : "TRUE", and a load job is currently running, the request will fail with this error.

• Failed to start new load for the source "source name". Max load task queue size limit breached. Limit is 64  (HTTP 400)

  Neptune supports queuing up as many as 64 loader jobs at a time. If an additional load request is submitted to the queue when it already contains 64 jobs, the request fails with this message.