The C++ reader is written completely from scratch. It does not need underlying java reader. It understands the ORC file specification and the format as such does not have anything to do with java. The metadata layer is all protobuf based which can be read in C++ as well.
Thanks Prasanth > On Aug 13, 2015, at 12:37 PM, Sean Luo <[email protected]> wrote: > > But won't C++ Orc reader need underlying Java? Or the current Apache Orc C++ > code does not work? > > > > On Thursday, August 13, 2015 12:34 PM, Prasanth J <[email protected]> > wrote: > > > Hi Sean > > We are still in the process of moving ORC java code out of Apache Hive > project. The java code will eventually land in the Apache Orc project. At > this point only the C++ reader for ORC is in Apache Orc master. > > Thanks > Prasanth > >> On Aug 13, 2015, at 12:27 PM, Sean Luo <[email protected] >> <mailto:[email protected]>> wrote: >> >> Hi community, >> >> I am new to Apache Orc. I have the impression that Orc was originally >> developed in Java, and I can see that there are Java code under Hive Orc: >> >> http://svn.apache.org/viewvc/hive/trunk/ql/src/java/org/apache/hadoop/hive/ql/io/orc/ >> >> <http://svn.apache.org/viewvc/hive/trunk/ql/src/java/org/apache/hadoop/hive/ql/io/orc/> >> >> But when I check out the Apache Orc git, I did not see any any Java code, >> and only C++: >> >> https://github.com/apache/orc/tree/master >> <https://github.com/apache/orc/tree/master> >> >> Can anyone clarify if the Orc is rewritten completely in C++, or the C++ >> code is just a wrapper on top of the old Java implementation? If the later, >> then why there is no Java code in Apache Orc git? >> >> Thanks a lot! >> >> Best, >> Sean > > >
