Spark：基于内容的推荐系统

ElasticSearch后台

spark索引创建
1.我们这里使用logstash上传到ES，分为三个字段分别是"当前商品、其他商品、相似度"（"current merchandise","other commodities","similarity"）
2.源数据格式

3.编写logstash配置文件

[root@master config]# cat spak-tj.conf 
input {
 file {
   path => ["/root/0920data.csv"]
   start_position => "beginning"
 }
}
filter {
  csv {
    separator => ","
    columns => ["current merchandise","other commodities","similarity"]
  } 
}
output {
  elasticsearch {
        hosts => ["master:9200"]
        index => "spark-tj"
  }
}

运行logstash程序：
进入到logstash主目录(/usr/local/logstash)运行：./bin/logstash -f config/spak-tj.conf
在ES中展示: