tencent cloud

1. 连接 Hive

登录 EMR 集群的 Master 节点，切换到 hadoop 用户并且进入 hive 目录：

[root@10 ~]# su hadoop
[hadoop@10 root]$ cd /usr/local/service/hive

2. 准备数据

创建数据文件（JSON 格式），编译以下内容并保存：

vim test.data
{"name":"Mary","age":12,"course":[{"name":"math","location":"b208"},{"name":"english","location":"b702"}],"grade":[99,98,95]}
{"name":"Bob","age":20,"course":[{"name":"music","location":"b108"},{"name":"history","location":"b711"}],"grade":[91,92,93]}

将数据文件存储在 hdfs 上：

hadoop fs -put ./test.data /

3. 创建表格

连接 Hive：

[hadoop@10 hive]$ hive

根据映射关系创建表格：

hive> CREATE TABLE test (name string, age int, course array<map<string,string>>, grade array<int>) ROW FORMAT SERDE 'org.apache.hive.hcatalog.data.JsonSerDe' STORED AS TEXTFILE;

4. 导入数据

hive>LOAD DATA INPATH '/test.data' into table test;

5. 检查数据是否导入成功

查询所有数据：

hive> select * from test;
OK
Mary    12    [{"name":"math","location":"b208"},{"name":"english","location":"b702"}]    [99,98,95]
Bob    20    [{"name":"music","location":"b108"},{"name":"history","location":"b711"}]   [91,92,93]
Time taken: 0.153 seconds, Fetched: 2 row(s)

查询每条记录的第一个得分：

hive> select grade[0] from test;
OK
99
91
Time taken: 0.374 seconds, Fetched: 2 row(s)

查询每条记录的第一个课程的名称和地点：

hive> select course[0]['name'], course[0]['location'] from test;
OK
math    b208
music    b108
Time taken: 0.162 seconds, Fetched: 2 row(s)