from ![Explorys][ExplorysBanner] [ExplorysBanner]: http://media.marketwire.com/attachments/201105/65384_ExplorysMedicalLogo_PMS.eps.jpg
Some highlights of Apothecary are:
- An alternative implementation of an InputFormat to allow you to use HBase as a source for mapreduce jobs.
- Reads directly from on disk storefiles
- In our testing, > 3 times faster
- Use responsibly, be aware of the operational constraints
- A drop in in-memory replacement for HTable
- Isolate your unit tests from your datastore without the pain
- Implements HTableInterface
- MapFile.Reader improvements
- Allows for multithreaded index lookups
- Isolate all of the Immutable Parameters from the disk operations
- Check out the project.
git clone git@github.com:ExplorysMedical/Apothecary.git- Edit the pom.xml to reflect the version of hadoop and hbase you are using. (Tested on CDH3U4)
- Install to your local repository from the project directory
mvn clean install- Add the following to your maven project:
<dependency>
<groupId>com.explorys.apothecary</groupId>
<artifactId>Apothecary</artifactID>
<version>1.0-SNAPSHOT</version>
</dependency>