mapreduce程序-awk脚本

mapper.awk

#!/bin/awk -f
BEGIN{
    RS = "\n";
	FS = " ";
	OFS = " ";   
}

{
	split($2, array_pname, ",");
	for(pname in array_pname) {
		array_counter[array_pname[pname]]++;
	}
}

END{
	for( pname in array_counter) {
		print pname, array_counter[pname];
	}
}

reducer.awk

#!/bin/awk -f
BEGIN{
    RS = "\n";
	FS = " ";
	OFS = " ";   
}

{
	array_counter[$1] += $2;
}

END{
	for( pname in array_counter ) {
		count = array_counter[pname];
		gsub(/\|/, " ", pname);
		print count, pname;
	}
}

hdpsubmit.sh

$HADOOP_HOME/bin/hadoop jar $HADOOP_HOME/contrib/streaming/hadoop-0.20.1.12-fb-streaming.jar \
    -D mapred.job.name=log_$yestoday \
    -D mapred.ignore.badcompress=true  \
    -D mapred.compress.map.output=true \
    -D mapred.map.output.compression.codec=org.apache.hadoop.io.compress.GzipCodec \
    -D mapred.output.compress=true \
    -D mapred.output.compression.codec=org.apache.hadoop.io.compress.GzipCodec \
    -D mapred.linerecordreader.maxlength=4096000 \
    -numReduceTasks 24 \
    -mapper $MAPRED_SCRIPT_DIR/mapper.awk   \
    -reducer $MAPRED_SCRIPT_DIR/reducer.awk \
    -file  $MAPRED_SCRIPT_DIR/mapper.awk    \
    -file  $MAPRED_SCRIPT_DIR/reducer.awk   \
    -input /home/hdp-zhushou/update_log/$yestoday/* \
    -output output_update_log/