MarkLogic Content Pump (mlcp) is a command-line tool that provides the fastest way to import, export, and copy data to or from MarkLogic databases. Core features of mlcp include:
- Bulk load billions of local files
- Split and load large, aggregate XML files or delimited text
- Bulk load billions of triples or quads from RDF files
- Archive and restore database contents across environments
- Export data from a database to a file system
- Copy subsets of data between databases
- Load documents from HDFS, including Hadoop SequenceFiles
You can run mlcp across many threads on a single machine or across many nodes in a Hadoop cluster.
The MarkLogic Connector for Hadoop is an extension to Hadoop’s MapReduce framework that allows you to easily and efficiently communicate with a MarkLogic database from within a Hadoop job. Use the Hadoop Connector to build Hadoop MapReduce jobs that interact with MarkLogic. The Hadoop Connector is also used by mlcp. Core features of the Hadoop Connector include:
- Process data in MarkLogic with Hadoop MapReduce for bulk analytics or transformation
- Persist data from Hadoop to MarkLogic for query and update
- Access MarkLogic text, geospatial, scalar, and document structure indexes to send only the most relevant data to Hadoop for processing
- Write results from MapReduce jobs to MarkLogic in parallel
- AWS Application Load Balancer support
- bug fixes
- Upgraded Jena library
- Bug fixes
- Support HDP 2.6
- Performance improvement in mlcp archive import and copy
- Bug fixes
- Performance optimization in mlcp archive import and mlcp copy
- Bug fixes
- Bug fixes
- New option to specify modules database and root directory for server-side transoformation
- Bug fixes
- New option to enable SSL communication
- New option to run behind a load balancer or firewall
- Survive server node failover events
- Server-side transform in batches
- Get and set document permission, collection and metadata in server-side transformation
- Performance improvements and bug fixes
For official product documentation, please refer to:
Wiki pages of this project contain useful information when you work on development:
- Required Software for the Hadoop Connector
- Required Software for mlcp
- Apache Maven (version >= 3.03) is required to build mlcp and the Hadoop Connector.
mlcp and Hadoop Connector can be built together. Steps to build:
$ git clone https://github.com/marklogic/marklogic-contentpump.git
$ cd marklogic-contentpump
$ mvn clean package -DskipTests=trueThe build writes to the respective deliverable directory under the top-level ./mlcp/ and ./mapreduce/ directories.
Alternatively, you can build mlcp and the Hadoop Connector independently from each component’s root directory (i.e. ./mlcp/ and ./mapreduce/) with the above command. Note that mlcp depends on the Hadoop Connector, so a successful build of the Hadoop Connector is required to build mlcp.
For information on contributing to this project see CONTRIBUTING.md. For information on working on development of this project see project wiki page.
The unit tests included in this repository are designed to provide illustrative examples of the APIs and to sanity check external contributions. MarkLogic Engineering runs a more comprehensive set of unit, integration, and performance tests internally. To run the unit tests, execute the following command from the marklogic-contentpump/ root directory:
$ mvn testFor detailed information about running unit tests, see Guideline to Running Tests.
If you have questions about mlcp or the Hadoop Connector, ask on StackOverflow. Tag your question with mlcp and marklogic. If you find a bug or would like to propose a new capability, file a GitHub issue.
mlcp and the Hadoop Connector are maintained by MarkLogic Engineering and distributed under the Apache 2.0 license. They are designed for use in production applications with MarkLogic Server. Everyone is encouraged to file bug reports, feature requests, and pull requests through GitHub. This input is critical and will be carefully considered. However, we can’t promise a specific resolution or timeframe for any request. In addition, MarkLogic provides technical support for release tags of mlcp and the Hadoop Connector to licensed customers under the terms outlined in the Support Handbook. For more information or to sign up for support, visit help.marklogic.com.