These notes are for Pig 0.6.0 release. Highlights ========== - Added Zebra as a contrib project. See http://wiki.apache.org/pig/zebra - Added UDFContext, gives UDFs a way to pass info from front to back end and gives UDFS access to JobConf in the backend. PIG-1085 - Added left outer join for fragment replicate join. PIG-1036 - Added ability to set job priority from Pig Latin. PIG-1025 - Enhanced multi-query to work with joins in some cases. PIG-983 - Reworked memory manager to significantly reduce GC Overhead and Out of Heap failures. PIG-975 - Added Accumulator interface for UDFs. PIG-979 - Over 100 bug fixes and improvements. Incompatibilities ================= PIG-922 changed LoadFunc.fieldsToRead's signature . Compiles LoadFuncs will work as is, since the Pig code checks if the method conforms to the new signature and only invokes it in the case that it does. If users wish to compile their LoadFunc, they will need to change this method to match the new signature. System Requirements =================== 1. Java 1.6.x or newer, preferably from Sun. Set JAVA_HOME to the root of your Java installation 2. Ant build tool: http://ant.apache.org - to build source only 3. Cygwin: http://www.cygwin.com/ - to run under Windows 4. This release is compatible with Hadoop 0.20.x releases Trying the Release ================== 1. Download pig-0.6.0.tar.gz 2. Unpack the file: tar -xzvf pig-0.6.0.tar.gz 3. Move into the installation directory: cd pig-0.6.0 4. To run pig without Hadoop cluster, execute the command below. This will take you into an interactive shell called grunt that allows you to navigate the local file system and execute Pig commands against the local files bin/pig -x local 5. To run on your Hadoop cluster, you need to set PIG_CLASSPATH environment variable to point to the directory with your hadoop-site.xml file and then run pig. The commands below will take you into an interactive shell called grunt that allows you to navigate Hadoop DFS and execute Pig commands against it export PIG_CLASSPATH=/hadoop/conf bin/pig 6. To build your own version of pig.jar run ant 7. To run unit tests run ant test 8. To build jar file with available user defined functions run commands below. This currently only works with Java 1.6.x. cd contrib/piggybank/java ant 9. To build the tutorial: cd tutorial ant 10. To run tutorial follow instructions in Relevant Documentation ====================== http://hadoop.apache.org/pig/docs/r0.6.0/ http://wiki.apache.org/pig/