big update: working iterating, pretty output, test scripts and more

1 files changed, 38 insertions, 0 deletions

diff --git a/dtrain/test/wc_pipes/wordcount.cc b/dtrain/test/wc_pipes/wordcount.cc
new file mode 100644
index 00000000..39560a31
--- /dev/null
+++ b/dtrain/test/wc_pipes/wordcount.cc
@@ -0,0 +1,38 @@
+#include "wordcount.hh"
+
+
+void
+WordcountMapper::map(HadoopPipes::MapContext & context)
+{
+  typedef boost::tokenizer<> tokenizer_t;
+  tokenizer_t tokenizer(context.getInputValue());
+
+  for( tokenizer_t::const_iterator i = tokenizer.begin();
+      tokenizer.end() != i; ++i ) {
+    context.emit(boost::to_lower_copy(*i), "1");
+  }
+}
+
+void
+WordcountReducer::reduce(HadoopPipes::ReduceContext & context)
+{
+  uint32_t count( 0 );
+
+  do {
+	++count;
+  } while( context.nextValue() );
+
+  std::cout << context.getInputKey() << endl;
+  context.emit( context.getInputKey(),
+                boost::lexical_cast<std::string>(count) );
+}
+
+
+int
+main( int argc, char * argv[] )
+{
+  HadoopPipes::TemplateFactory2<WordcountMapper,
+                                WordcountReducer> factory;
+  return HadoopPipes::runTask( factory );
+}
+