Hive 配置单元排序依据在子查询中不起作用

我有这个疑问-

CREATE TABLE traces_orc1 
(pc BIGINT, 
 opcode STRING, 
 frequency STRING) 
ROW FORMAT DELIMITED FIELDS TERMINATED BY '44' 
STORED AS ORC;

INSERT OVERWRITE TABLE traces_orc1 
SELECT CONV(pc,16,10) as pc, 
opcode, COUNT(*) 
FROM traces_texttemp as frequency 
GROUP BY pc, opcode;


CREATE TABLE traces_orctemp 
(pc BIGINT, 
 opcode STRING, 
 frequency STRING) 
ROW FORMAT DELIMITED FIELDS TERMINATED BY '44' 
STORED AS ORC;

INSERT OVERWRITE TABLE traces_orctemp 
SELECT pc, opcode, frequency FROM traces_orc1 ORDER BY pc ASC;

我试图把这两个语句合并成一个-

CREATE TABLE traces_orc2 (
pcount BIGINT, 
opcode STRING, 
frequency STRING) 
ROW FORMAT DELIMITED FIELDS TERMINATED BY '44' 
STORED AS TEXTFILE;

INSERT OVERWRITE TABLE traces_orc2 
SELECT CONV(TRIM(traces_text.pc),16,10) as pcount, 
opcode, COUNT(*) as frequency 
FROM traces_text
GROUP BY pcount,opcode
ORDER BY pcount ASC;

但这行不通。当我将其作为一条语句编写时，

orderby

不起作用

任何输入？？？

您是否尝试过：

INSERT OVERWRITE TABLE traces_orc2 
select pcount, opcode, frequency from 
(SELECT CONV(TRIM(traces_text.pc),16,10) as pcount, 
opcode, COUNT(*) as frequency 
FROM traces_text
GROUP BY pcount,opcode)
ORDER BY pcount ASC;

---

它可能会强制Hive启动第二个减少阶段进行排序。

您在第一次覆盖中有

跟踪\u text temp

，在第三次覆盖中有

跟踪\u text

。这些是从哪里来的？这只是我加载数据的两个表