Hive数据操作语言（DML)

1. 加载数据（load)

• LOAD DATA [LOCAL] INPATH ‘filepath’ [OVERWRITE] INTO TABLE tablename [PARTITION (partcol1=val1, partcol2=val2 …)]

  加载数据到表中时，Hive不做任何转换。加载操作只是把数据拷贝或移动操作，即移动数据文件到Hive表相应的位置。
  加载的目标可以是一个表，也可以是一个分区。如果表是分区的，则必须通过指定所有分区列的值来指定一个表的分区。
  filepath可以是一个文件，也可以是一个目录。不管什么情况下，filepath被认为是一个文件集合。

• LOCAL：表示输入文件在本地文件系统（Linux），如果没有加LOCAL，hive则会去HDFS上查找该文件。

• OVERWRITE：重写，覆盖。
• PARTITION：如果表中存在分区，可以按照分区进行导入。

导入数据

• 建表

      1. hive> CREATE TABLE emp (
      2.     > empno int,
      3.     > ename string,
      4.     > job string,
      5.     > mgr int,
      6.     > hiredate string,
      7.     > salary double,
      8.     > comm double,
      9.     > deptno int
      10.     > ) ROW FORMAT DELIMITED FIELDS TERMINATED BY "\n";
      11. OK
      12. Time taken: 0.54 seconds
  

• 导入本地文本

      1.  LOAD DATA LOCAL INPATH '/home/hadoop/emp.txt' OVERWRITE INTO TABLE emp;
  
      2. hive> select * from emp;
      3. OK
      4. 7369    SMITH   CLERK   7902    1980/12/17      800.0   NULL    20
      5. 7499    ALLEN   SALESMAN        7698    1981/2/20       1600.0  300.0   30
      6. 7521    WARD    SALESMAN        7698    1981/2/22       1250.0  500.0   30
      7. 7566    JONES   MANAGER 7839    1981/4/2        2975.0  NULL    20
      8. 7654    MARTIN  SALESMAN        7698    1981/9/28       1250.0  1400.0  30
      9. 7698    BLAKE   MANAGER 7839    1981/5/1        2850.0  NULL    30
      10. 7782    CLARK   MANAGER 7839    1981/6/9        2450.0  NULL    10
      11. 7788    SCOTT   ANALYST 7566    1987/4/19       3000.0  NULL    20
      12. 7839    KING    PRESIDENT       NULL    1981/11/17      5000.0  NULL    10
      13. 7844    TURNER  SALESMAN        7698    1981/9/8        1500.0  0.0     30
      14. 7876    ADAMS   CLERK   7788    1987/5/23       1100.0  NULL    20
      15. 7900    JAMES   CLERK   7698    1981/12/3       950.0   NULL    30
      16. 7902    FORD    ANALYST 7566    1981/12/3       3000.0  NULL    20
      17. 7934    MILLER  CLERK   7782    1982/1/23       1300.0  NULL    10
      18. Time taken: 0.938 seconds, Fetched: 14 row(s)
  

• 导入分区

      1. hive> load data local inpath '/home/hadoop/dept.txt' into table dept partition (dt='2018-09-09');
      2. Loading data to table default.dept partition (dt=2018-09-09)
      3. Partition default.dept{dt=2018-09-09} stats: [numFiles=1, totalSize=84]
      4. OK
      5. Time taken: 10.631 seconds
      6. hive> select * form dept;
      7. FAILED: ParseException line 1:9 missing EOF at 'form' near '*'
      8. hive> select * from dept;
      9. OK
      10. 10      ACCOUNTING      NEW YORK        2018-08-08
      12. 20      RESEARCH        DALLAS  2018-08-08
      13. 30      SALES           CHICAGO 2018-08-08
      14. 40      OPERATIONS      BOSTON  2018-08-08
      15. 10      ACCOUNTING      NEW YORK        2018-09-09
      16. 20      RESEARCH        DALLAS  2018-09-09
      17. 30      SALES           CHICAGO 2018-09-09
      18. 40      OPERATIONS      BOSTON  2018-09-09
      19. Time taken: 1.385 seconds, Fetched: 8 row(s)
  

2. 插入数据（insert into)

标准插入

• INSERT OVERWRITE TABLE tablename1 [PARTITION (partcol1=val1, partcol2=val2 …) [IF NOT EXISTS]] select_statement1 FROM from_statement;

  INSERT INTO TABLE tablename1 [PARTITION (partcol1=val1, partcol2=val2 …)] select_statement1 FROM from_statement;
  可以使用PARTITION 关键字，进行分区插入。
  OVERWRITE是否选择覆盖。
  使用插入语法会跑mr作业。

• 向emp1表中插入emp表

      1. hive> insert overwrite table emp1 select * from emp;
      2. Query ID = hadoop_20180109081212_d62e58f3-946c-465e-999d-2ddf0d76d807
      3. Total jobs = 3
      4. Launching Job 1 out of 3
  
      5. hive> select * from emp1;
      6. OK
      7. 7369    SMITH   CLERK   7902    1980/12/17      800.0   NULL    20
      8. 7499    ALLEN   SALESMAN        7698    1981/2/20       1600.0  300.0   30
      9. 7521    WARD    SALESMAN        7698    1981/2/22       1250.0  500.0   30
      10. 7566    JONES   MANAGER 7839    1981/4/2        2975.0  NULL    20
      11. 7654    MARTIN  SALESMAN        7698    1981/9/28       1250.0  1400.0  30
      12. 7698    BLAKE   MANAGER 7839    1981/5/1        2850.0  NULL    30
      13. 7782    CLARK   MANAGER 7839    1981/6/9        2450.0  NULL    10
      14. 7788    SCOTT   ANALYST 7566    1987/4/19       3000.0  NULL    20
      15. 7839    KING    PRESIDENT       NULL    1981/11/17      5000.0  NULL    10
      16. 7844    TURNER  SALESMAN        7698    1981/9/8        1500.0  0.0     30
      17. 7876    ADAMS   CLERK   7788    1987/5/23       1100.0  NULL    20
      18. 7900    JAMES   CLERK   7698    1981/12/3       950.0   NULL    30
      19. 7902    FORD    ANALYST 7566    1981/12/3       3000.0  NULL    20
      20. 7934    MILLER  CLERK   7782    1982/1/23       1300.0  NULL    10
      21. Time taken: 0.211 seconds, Fetched: 14 row(s)
  

• 按字段进行插入时，不要把字段顺序写错，否则插入时不会报错，但要查找数据时，数据查询不到
  演示：
  把job,ename顺序写错进行插入

      1. hive> insert overwrite table emp2 select empno,job,ename,mgr,hiredate,salary,comm,deptno from emp;
  

  再插入emp表

      1.  hive> insert into table emp2 select * from emp;
      2. hive> select * from emp2;
      3. OK
      4. 7369    CLERK   SMITH   7902    1980/12/17      800.0   NULL    20
      5. 7499    SALESMAN        ALLEN   7698    1981/2/20       1600.0  300.0   30
      6. 7521    SALESMAN        WARD    7698    1981/2/22       1250.0  500.0   30
      7. 7566    MANAGER JONES   7839    1981/4/2        2975.0  NULL    20
      8. 7654    SALESMAN        MARTIN  7698    1981/9/28       1250.0  1400.0  30
      9. 7698    MANAGER BLAKE   7839    1981/5/1        2850.0  NULL    30
      10. 7782    MANAGER CLARK   7839    1981/6/9        2450.0  NULL    10
      11. 7788    ANALYST SCOTT   7566    1987/4/19       3000.0  NULL    20
      12. 7839    PRESIDENT       KING    NULL    1981/11/17      5000.0  NULL    10
      13. 7844    SALESMAN        TURNER  7698    1981/9/8        1500.0  0.0     30
      14. 7876    CLERK   ADAMS   7788    1987/5/23       1100.0  NULL    20
      15. 7900    CLERK   JAMES   7698    1981/12/3       950.0   NULL    30
      16. 7902    ANALYST FORD    7566    1981/12/3       3000.0  NULL    20
      17. 7934    CLERK   MILLER  7782    1982/1/23       1300.0  NULL    10
      18. 7369    SMITH   CLERK   7902    1980/12/17      800.0   NULL    20
      19. 7499    ALLEN   SALESMAN        7698    1981/2/20       1600.0  300.0   30
      20. 7521    WARD    SALESMAN        7698    1981/2/22       1250.0  500.0   30
      21. 7566    JONES   MANAGER 7839    1981/4/2        2975.0  NULL    20
      22. 7654    MARTIN  SALESMAN        7698    1981/9/28       1250.0  1400.0  30
      23. 7698    BLAKE   MANAGER 7839    1981/5/1        2850.0  NULL    30
      24. 7782    CLARK   MANAGER 7839    1981/6/9        2450.0  NULL    10
      25. 7788    SCOTT   ANALYST 7566    1987/4/19       3000.0  NULL    20
      26. 7839    KING    PRESIDENT       NULL    1981/11/17      5000.0  NULL    10
      27. 7844    TURNER  SALESMAN        7698    1981/9/8        1500.0  0.0     30
      28. 7876    ADAMS   CLERK   7788    1987/5/23       1100.0  NULL    20
      29. 7900    JAMES   CLERK   7698    1981/12/3       950.0   NULL    30
      30. 7902    FORD    ANALYST 7566    1981/12/3       3000.0  NULL    20
      31. 7934    MILLER  CLERK   7782    1982/1/23       1300.0  NULL    10
      32. Time taken: 2.363 seconds, Fetched: 28 row(s)
  

  查询结果前14条记录job,ename是反的，可以成功插入，但是在查询相关数据时查询不到结果

多行插入

• FROM from_statement
  INSERT OVERWRITE TABLE tablename1 [PARTITION (partcol1=val1, partcol2=val2 …) [IF NOT EXISTS]] select_statement1
  [INSERT OVERWRITE TABLE tablename2 [PARTITION … [IF NOT EXISTS]] select_statement2]
  [INSERT INTO TABLE tablename2 [PARTITION …] select_statement2] …;

多行插入是把from提到语句首，其实质就是简化标准插入

手动插入数据

• INSERT INTO TABLE tablename [PARTITION (partcol1[=val1], partcol2[=val2] …)] VALUES values_row [, values_row …]

      1. e>  create table stu(
      2.  > id int,
      3.  > name string
      4.  > )
      5. ROW FORMAT DELIMITED FIELDS TERMINATED BY '\t';
      6. OK
      7. Time taken: 0.405 seconds
      8. hive> select * from stu;
      9. OK
  
      10. hive> insert into table stu values(1,'zhangsan'),(2,'lisi);
  
      11. hive> select * from stu;
      12. OK
      13. 1       zhangsan
      14. 2       lisi
  

3. 数据导出

• 标准导出
  INSERT OVERWRITE [LOCAL] DIRECTORY directory1
  [ROW FORMAT row_format] [STORED AS file_format] (Note: Only available starting with Hive 0.11.0)
  SELECT … FROM …

  LOCAL：加上LOCAL关键字代表导入本地系统，不加默认导入HDFS；
  STORED AS:可以指定存储格式。

      1. hive> insert overwrite local directory '/home/hadoop/data' ROW FORMAT DELIMITED FIELDS TERMINATED BY '\t' 
      select * from stu;
  
      2. [hadoop@zydatahadoop001 data]$ pwd
      3. /home/hadoop/data
  
      4. [hadoop@zydatahadoop001 data]$ cat 000000_0 
      5. 1       zhangsan
      6. 2       lisi
  

• 多条导出
  FROM from_statement
  INSERT OVERWRITE [LOCAL] DIRECTORY directory1 select_statement1
  [INSERT OVERWRITE [LOCAL] DIRECTORY directory2 select_statement2] …

      1. hive> from emp
      2.     > INSERT OVERWRITE  LOCAL DIRECTORY '/home/hadoop/tmp/hivetmp1'
      3.     > ROW FORMAT DELIMITED FIELDS TERMINATED BY "\t"
      4.     > select empno, ename  
      5.     > INSERT OVERWRITE  LOCAL DIRECTORY '/home/hadoop/tmp/hivetmp2'
      6.     > ROW FORMAT DELIMITED FIELDS TERMINATED BY "\t"
      7.     > select ename;
  
      8. [hadoop@zydatahadoop001 tmp]$ pwd
      9. /home/hadoop/tmp
      10. [hadoop@zydatahadoop001 tmp]$ cat hivetmp1/000000_0 
      11. 7369    SMITH
      12. 7499    ALLEN
      13. 7521    WARD
      14. 7566    JONES
      15. 7654    MARTIN
      16. 7698    BLAKE
      17. 7782    CLARK
      18. 7788    SCOTT
      19. 7839    KING
      20. 7844    TURNER
      21. 7876    ADAMS
      22. 7900    JAMES
      23. 7902    FORD
      24. 7934    MILLER
      25. [hadoop@zydatahadoop001 tmp]$ cat hivetmp2/000000_0 
      26. SMITH
      27. ALLEN
      28. WARD
      29. JONES
      30. MARTIN
      31. BLAKE
      32. CLARK
      33. SCOTT
      34. KING
      35. TURNER
      36. ADAMS
      37. JAMES
      38. FORD
      39. MILLER
  

4. SELECT

where条件语句

• 查询员工表deptno=10的员工

      1. hive> select * from emp where deptno=10;
      2. OK
      3. 7782    CLARK   MANAGER 7839    1981/6/9        2450.0  NULL    10
      4. 7839    KING    PRESIDENT       NULL    1981/11/17      5000.0  NULL    10
      5. 7934    MILLER  CLERK   7782    1982/1/23       1300.0  NULL    10
      6. Time taken: 1.144 seconds, Fetched: 3 row(s)
  

• 查询员工编号小于等于7800的员工

      1. hive> select * from emp where empno <= 7800;
  

• 查询员工工资大于1000小于1500的员工

      1. hive> select * from emp where salary between 1000 and 1500;
  

• 查询前5条记录

      1. hive> select * from emp limit 5;
  

• 查询编号为7566或7499的员工

      1. hive> select * from emp where empno in(7566,7499);
  

• 查询有津贴不为空的员工

      1. hive> select * from emp where comm is not null;
  

来自@若泽大数据