Skip to content

G
gic-spark-tag-4.0
Commit f5b3e5fb by guos
Options

会员标签4.0

parent 1800bc4c
Showing with 38 additions and 37 deletions
src/main/java/com/gic/spark/filter/TagAssociatedPurchaseRateFilter.java
......@@ -6,6 +6,7 @@ import com.gic.spark.entity.bean.TrdEcuBrandLabelBean;
import com.gic.spark.entity.request.AbstractFilterRequest;
import com.gic.spark.entity.request.TagConsumeDoubleRequest;
import com.gic.spark.entity.request.TagConsumeRequest;
import com.gic.spark.util.CommonUtil;
import org.apache.spark.api.java.JavaRDD;
import scala.Tuple2;
......@@ -40,18 +41,17 @@ public class TagAssociatedPurchaseRateFilter extends AbstractTagConsumFilter{
TagConsumeDoubleRequest consumeRequest=(TagConsumeDoubleRequest)request;
JavaRDD<TrdEcuBrandLabelBean> consumeRDD= MysqlRddManager.getPojoFromDataset(dataSourceHive.getDatasetByEntId(enterpriseId), TrdEcuBrandLabelBean.class).javaRDD();
consumeRDD=statisticsTypeHandle(consumeRDD,consumeRequest);
JavaRDD<Long>ecuRdd=consumeRDD.mapToPair(data-> Tuple2.apply(data.getEcu_id(),data)).groupByKey()
JavaRDD<Long>ecuRdd=consumeRDD.mapToPair(data-> Tuple2.apply(data.getEcu_id(),data))
.reduceByKey((x,y)->{
x.setSeff_goods_num(x.getSeff_goods_num()+y.getSeff_goods_num());
x.setOrder_times(x.getOrder_times()+y.getOrder_times());
return x;
})
.mapPartitions(data->{
List<Long> result=new ArrayList();
while (data.hasNext()){
Tuple2<Long,Iterable<TrdEcuBrandLabelBean>> tp2=data.next();
long totalGodsNum=0;
long totalOrderTimes=0;
for(TrdEcuBrandLabelBean consumeBean:tp2._2()){
totalGodsNum+=consumeBean.getSeff_goods_num();
totalOrderTimes+=consumeBean.getOrder_times();
}
double jointRate=totalGodsNum/totalOrderTimes;
Tuple2<Long,TrdEcuBrandLabelBean> tp2=data.next();
double jointRate= CommonUtil.isEmptyInteger2int(tp2._2().getSeff_goods_num())/CommonUtil.isEmptyInteger2int(tp2._2().getOrder_times());
switch (consumeRequest.getNumberType()){
case gt:
if(jointRate>consumeRequest.getBeginNum()){
......
src/main/java/com/gic/spark/filter/TagAverageDiscountFactorFilter.java
......@@ -6,6 +6,7 @@ import com.gic.spark.entity.bean.TrdEcuBrandLabelBean;
import com.gic.spark.entity.request.AbstractFilterRequest;
import com.gic.spark.entity.request.TagConsumeDoubleRequest;
import com.gic.spark.entity.request.TagConsumeRequest;
import com.gic.spark.util.CommonUtil;
import org.apache.spark.api.java.JavaRDD;
import scala.Tuple2;
......@@ -42,18 +43,17 @@ public class TagAverageDiscountFactorFilter extends AbstractTagConsumFilter {
TagConsumeDoubleRequest consumeRequest=(TagConsumeDoubleRequest)request;
JavaRDD<TrdEcuBrandLabelBean> consumeRDD= MysqlRddManager.getPojoFromDataset(dataSourceHive.getDatasetByEntId(enterpriseId), TrdEcuBrandLabelBean.class).javaRDD();
consumeRDD=statisticsTypeHandle(consumeRDD,consumeRequest);
JavaRDD<Long>ecuRdd=consumeRDD.mapToPair(data-> Tuple2.apply(data.getEcu_id(),data)).groupByKey()
JavaRDD<Long>ecuRdd=consumeRDD.mapToPair(data-> Tuple2.apply(data.getEcu_id(),data))
.reduceByKey((x,y)->{
x.setPay_amt(x.getPay_amt()+y.getPay_amt());
x.setTotal_amt(x.getTotal_amt()+y.getTotal_amt());
return x;
})
.mapPartitions(data->{
List<Long>result=new ArrayList();
while (data.hasNext()){
double payAmt=0;
double totalAmt=0;
Tuple2<Long,Iterable<TrdEcuBrandLabelBean>> tp2=data.next();
for(TrdEcuBrandLabelBean consumeBean:tp2._2()){
payAmt=consumeBean.getPay_amt();
totalAmt=consumeBean.getTotal_amt();
}
double avgDiscountRate=payAmt/totalAmt;
Tuple2<Long,TrdEcuBrandLabelBean> tp2=data.next();
double avgDiscountRate= CommonUtil.isEmptyDouble2double(tp2._2().getPay_amt())/CommonUtil.isEmptyDouble2double(tp2._2().getTotal_amt());
switch (consumeRequest.getNumberType()){
case gt:
if(avgDiscountRate>consumeRequest.getBeginNum()){
......
src/main/java/com/gic/spark/filter/TagConsumptionSleepDaysFilter.java
......@@ -41,47 +41,48 @@ public class TagConsumptionSleepDaysFilter extends AbstractTagConsumFilter{
JavaRDD<TrdEcuBrandLabelBean> consumeRDD= MysqlRddManager.getPojoFromDataset(dataSourceHive.getDatasetByEntId(enterpriseId), TrdEcuBrandLabelBean.class).javaRDD();
consumeRDD=statisticsTypeHandle(consumeRDD,consumeRequest);
JavaRDD<Long>ecuRdd=consumeRDD.mapToPair(data-> Tuple2.apply(data.getEcu_id(),data.getSleep_days()))
.reduceByKey((x,y)->x>y?y:x)
.mapPartitions(data->{
List<Long>result=new ArrayList();
.mapPartitionsToPair(data->{
List<Tuple2<Long,Integer>>result=new ArrayList();
while (data.hasNext()){
Tuple2<Long,Integer>tp2=data.next();
switch (consumeRequest.getNumberType()){
case gt:
if(tp2._2()>consumeRequest.getBeginNum()){
result.add(tp2._1());
result.add(tp2);
}
break;
case gte:
if(tp2._2()>=consumeRequest.getBeginNum()){
result.add(tp2._1());
result.add(tp2);
}
break;
case lt:
if(tp2._2()<consumeRequest.getEndNum()){
result.add(tp2._1());
result.add(tp2);
}
break;
case lte:
if(tp2._2()<=consumeRequest.getEndNum()){
result.add(tp2._1());
result.add(tp2);
}
break;
case eq:
if(tp2._2()==consumeRequest.getEqualNum()){
result.add(tp2._1());
result.add(tp2);
}
break;
case between:
if(tp2._2()>=consumeRequest.getBeginNum()
&&tp2._2()<=consumeRequest.getEndNum()){
result.add(tp2._1());
result.add(tp2);
}
default:break;
}
}
return result.iterator();
});
})
.reduceByKey((x,y)->x)
.map(data->data._1());
return ecuRdd;
}
}
src/main/java/com/gic/spark/filter/TagPerCustomerTransactionFilter.java
......@@ -5,6 +5,7 @@ import com.gic.spark.datasource.mysql.MysqlRddManager;
import com.gic.spark.entity.bean.TrdEcuBrandLabelBean;
import com.gic.spark.entity.request.AbstractFilterRequest;
import com.gic.spark.entity.request.TagConsumeRequest;
import com.gic.spark.util.CommonUtil;
import org.apache.spark.api.java.JavaRDD;
import scala.Tuple2;
......@@ -39,18 +40,17 @@ public class TagPerCustomerTransactionFilter extends AbstractTagConsumFilter{
TagConsumeRequest consumeRequest=(TagConsumeRequest)request;
JavaRDD<TrdEcuBrandLabelBean> consumeRDD= MysqlRddManager.getPojoFromDataset(dataSourceHive.getDatasetByEntId(enterpriseId), TrdEcuBrandLabelBean.class).javaRDD();
consumeRDD=statisticsTypeHandle(consumeRDD,consumeRequest);
JavaRDD<Long>ecuRdd=consumeRDD.mapToPair(data-> Tuple2.apply(data.getEcu_id(),data)).groupByKey()
JavaRDD<Long>ecuRdd=consumeRDD.mapToPair(data-> Tuple2.apply(data.getEcu_id(),data))
.reduceByKey((x,y)->{
x.setPay_amt(x.getPay_amt()+y.getPay_amt());
x.setOrder_times(x.getOrder_times()+y.getOrder_times());
return x;
})
.mapPartitions(data->{
List<Long> result=new ArrayList();
while (data.hasNext()){
Tuple2<Long,Iterable<TrdEcuBrandLabelBean>> tp2=data.next();
int consumeTimes=0;
double payAmt=0;
for(TrdEcuBrandLabelBean consumeBean:tp2._2()){
consumeTimes+=consumeBean.getOrder_times();
payAmt+=consumeBean.getPay_amt();
}
double CusSinglePiece=payAmt/consumeTimes;
Tuple2<Long,TrdEcuBrandLabelBean> tp2=data.next();
double CusSinglePiece= CommonUtil.isEmptyDouble2double(tp2._2().getPay_amt())/CommonUtil.isEmptyInteger2int(tp2._2().getOrder_times());
switch (consumeRequest.getNumberType()){
case gt:
if(CusSinglePiece>consumeRequest.getBeginNum()){
......
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment