move AWS billing data to hive partitioned paths

gistfile1.txt

package com.bizo.aws.billing
import com.amazonaws.services.s3._
import com.bizo.util.aws.SystemPropertyOrEnvAWSCredentials
import com.amazonaws.services.s3.model.ListObjectsRequest
import scala.collection.JavaConversions._

object BillingMover {
  val bucket = ...

  def main(args: Array[String]) {
    val s3 = new AmazonS3Client(new SystemPropertyOrEnvAWSCredentials)
    
    val listing = s3.listObjects(new ListObjectsRequest().withBucketName(bucket).withDelimiter("/"))
    
    val billPattern = """.*-(\d{4})-(\d{2})\.csv""".r
    for (o <- listing.getObjectSummaries) {
      
      o.getKey match {
        case billPattern(year, month) => {
          println(o.getKey + ", " + year + ", " + month)
          val partitioned = "partitioned/year=%s/month=%s/%s".format(year, month, o.getKey)
          
          s3.copyObject(bucket, o.getKey, bucket, partitioned)
          s3.deleteObject(bucket, o.getKey)
        }
        case _ => // ignore
      }
    }
  }
}