with t1 as (

    select

        imei

        ,src_pkg

        ,src_type

        ,app_version_name

    from bi_quickgame.dw_qgcrpk_boot_di

    where day = '${etl_date}'

    group by

        imei, src_pkg, src_type, app_version_name

)

EXPLAIN

with t as(

select regexp_replace(reflect("java.util.UUID", "randomUUID"), "-", "") AS id --生成一个随机id

,'zengkun' as name

)

select * from t

union all

select * from t

;

Explain

STAGE DEPENDENCIES:

  Stage-1 is a root stage

  Stage-0 depends on stages: Stage-1

STAGE PLANS:

  Stage: Stage-1

    Map Reduce

      Map Operator Tree:

          TableScan

            alias: _dummy_table

            Row Limit Per Split: 1

            Statistics: Num rows: 1 Data size: 1 Basic stats: COMPLETE Column stats: COMPLETE

            Select Operator

              expressions: regexp_replace(reflect('java.util.UUID','randomUUID'), '-', '') (type: string), 'zengkun' (type: string)

              outputColumnNames: _col0, _col1

              Statistics: Num rows: 1 Data size: 1 Basic stats: COMPLETE Column stats: COMPLETE

              Union

                Statistics: Num rows: 2 Data size: 2 Basic stats: COMPLETE Column stats: COMPLETE

                File Output Operator

                  compressed: false

                  Statistics: Num rows: 2 Data size: 2 Basic stats: COMPLETE Column stats: COMPLETE

                  table:

                      input format: org.apache.hadoop.mapred.TextInputFormat

                      output format: org.apache.hadoop.hive.ql.io.HiveIgnoreKeyTextOutputFormat

                      serde: org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe

          TableScan

            alias: _dummy_table

            Row Limit Per Split: 1

            Statistics: Num rows: 1 Data size: 1 Basic stats: COMPLETE Column stats: COMPLETE

            Select Operator

              expressions: regexp_replace(reflect('java.util.UUID','randomUUID'), '-', '') (type: string), 'zengkun' (type: string)

              outputColumnNames: _col0, _col1

              Statistics: Num rows: 1 Data size: 1 Basic stats: COMPLETE Column stats: COMPLETE

              Union

                Statistics: Num rows: 2 Data size: 2 Basic stats: COMPLETE Column stats: COMPLETE

                File Output Operator

                  compressed: false

                  Statistics: Num rows: 2 Data size: 2 Basic stats: COMPLETE Column stats: COMPLETE

                  table:

                      input format: org.apache.hadoop.mapred.TextInputFormat

                      output format: org.apache.hadoop.hive.ql.io.HiveIgnoreKeyTextOutputFormat

                      serde: org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe

  Stage: Stage-0

    Fetch Operator

      limit: -1

      Processor Tree:

        ListSink

-- 获取所有除曝光的所有埋点

create temporary table bi_gamecenter_dev.tmp_da_gc_origin_event_di_1 as

select

    t.*

from bi_gamecenter_dev.dw_gc_oprt_log_h t

where day  = '${etl_date}'

and not (  --delete the exposure data

     action = '/clientRequest/userOperationLog'

     and (

        fields['exposure_stat'] is not null

        or fields['appexpo'] is not null

        or fields['search_exposure'] is not null

        or fields['exposure'] is not null

     )

);

-- 前天排行榜访问数据

drop table IF EXISTS bi_gamecenter_dev.tmp_gc_board_exposure_1_${etl_datekey};

create table bi_gamecenter_dev.tmp_gc_board_exposure_1_${etl_datekey} as

select --曝光

    origin_event as origin,

    app_version_code,

    imei

from

    bi_gamecenter_dev.dw_gc_exposure_log_d

where day = date_sub('${etl_date}',1) 

and   origin_event in ('530','531','532','533')

-- 过风控

and risk_flag = 0

;

drop table IF EXISTS bi_gamecenter_dev.tmp_gc_board_exposure_1_${etl_datekey};

1.使用方便，无需关注结果数据内容。

2.写法简易，不需要建表，直接使用即可

3.可读性强

1.如果多次使用，会多次scan表，降低效率

2.数据自查的时候，不容易定位问题

1.使用较为方便，单个进程中存在，自动删除。

2.具备复用性，在相同任务中，重复使用结果。

1.任务结束既删除，不方便过程数据校验。

2.无法创建分区表，只能简单创建临时表。

1.直接创建真实表，并且落地hdfs，方便过程数据校验。

2.可以创建分区临时表等操作，丰富临时表。

1.使用不方便，每次需要先drop再create，完成后再drop。