数据背景

hive (ods)> select * from jsont1;
OK
jsont1.username	jsont1.age	jsont1.sex	jsont1.json
user1	18	male	{"id":1,"ids":[101,102,103],"total_number":3}
user2	28	female	{"id":2,"ids":[201,202,203],"total_number":3}
user3	38	male	{"id":3,"ids":[301,302,303,304],"total_number":4}
user4	48	female	{"id":4,"ids":[401,402,403,404,405],"total_number":5}
user5	58	male	{"id":5,"ids":[501,502,503],"total_number":3}
user6	68	female	{"id":6,"ids":[601,602,603],"total_number":3}
user7	78	male	{"id":7,"ids":[701,702,703],"total_number":3}
	NULL	NULL	NULL
Time taken: 0.933 seconds, Fetched: 8 row(s)

一.使用hive内置函数处理

1:通过get_json_object()单个取

hive (ods)> select username,age,sex,get_json_object(json,"$.id") id ,get_json_object(json,"$.ids") ids ,get_json_object(json,"$.total_number") totalNumber
          > from jsont1;
OK
username	age	sex	id	ids	totalnumber
user1	18	male	1	[101,102,103]	3
user2	28	female	2	[201,202,203]	3
user3	38	male	3	[301,302,303,304]	4
user4	48	female	4	[401,402,403,404,405]	5
user5	58	male	5	[501,502,503]	3
user6	68	female	6	[601,602,603]	3
user7	78	male	7	[701,702,703]	3
	NULL	NULL	NULL	NULL	NULL
Time taken: 0.562 seconds, Fetched: 8 row(s)

还可以取数组中的某个元素

hive (ods)> select username,age,sex,get_json_object(json,"$.id") id ,
          > get_json_object(json,"$.ids[0]") ids0,
          > get_json_object(json,"$.ids[1]") ids1,
          > get_json_object(json,"$.ids[2]") ids2,
          > get_json_object(json,"$.ids[3]") ids3,
          > get_json_object(json,"$.ids[4]") ids4,
          > get_json_object(json,"$.ids[5]") ids5,
          > get_json_object(json,"$.total_number") totalNumber
          > from jsont1;
OK
username	age	sex	id	ids0	ids1	ids2	ids3	ids4	ids5	totalnumber
user1	18	male	1	101	102	103	NULL	NULL	NULL	3
user2	28	female	2	201	202	203	NULL	NULL	NULL	3
user3	38	male	3	301	302	303	304	NULL	NULL	4
user4	48	female	4	401	402	403	404	405	NULL	5
user5	58	male	5	501	502	503	NULL	NULL	NULL	3
user6	68	female	6	601	602	603	NULL	NULL	NULL	3
user7	78	male	7	701	702	703	NULL	NULL	NULL	3
	NULL	NULL	NULL	NULL	NULL	NULL	NULL	NULL	NULL	NULL
Time taken: 6.033 seconds, Fetched: 8 row(s)
hive (ods)>

2.通过hive内置函数json_tuple()一起取

hive (ods)> select json_tuple(json,'id','ids','total_number')
          > from jsont1;
OK
c0	c1	c2
1	[101,102,103]	3
2	[201,202,203]	3
3	[301,302,303,304]	4
4	[401,402,403,404,405]	5
5	[501,502,503]	3
6	[601,602,603]	3
7	[701,702,703]	3
NULL	NULL	NULL
Time taken: 1.076 seconds, Fetched: 8 row(s)

以上处理方式缺点:
(1).通过hive内置函数解析出的json都为json字符串并非我们想要的字符数组.
我们仔细观察它们俩者之间的不同

"[101,102,103]"-->["101","102","103"]
如果我们要通过hql处理如下：
hive (ods)> select regexp_replace("[101,102,103]","\\[|\\]","");
OK
_c0
101,102,103
Time taken: 0.237 seconds, Fetched: 1 row(s)
hive (ods)> select split(regexp_replace("[101,102,103]","\\[|\\]",""),",");
OK
_c0
["101","102","103"]

(2)通过json_tuple()进行都分个取出，但是发现不能和原字段并列显示

hive (ods)> select username,age,sex,json_tuple(json,'id','ids','total_number')
          > from jsont1;
FAILED: SemanticException [Error 10081]: UDTF's are not supported outside the SELECT clause, nor nested in expressions

ps:报错:我自己猜测是hive中没有json数组类型
补救措施 lateral view视图
通过下列方式可以(lateral view 拼接视图)

hive (ods)> select username,age,sex,id,ids,num
          > from jsont1
          > lateral view json_tuple(json,'id','ids','total_number') t1 as id,ids,num;
OK
username	age	sex	id	ids	num
user1	18	male	1	[101,102,103]	3
user2	28	female	2	[201,202,203]	3
user3	38	male	3	[301,302,303,304]	4
user4	48	female	4	[401,402,403,404,405]	5
user5	58	male	5	[501,502,503]	3
user6	68	female	6	[601,602,603]	3
user7	78	male	7	[701,702,703]	3
	NULL	NULL	NULL	NULL	NULL
Time taken: 0.64 seconds, Fetched: 8 row(s)

(3)如果我们需要对字符数组炸裂做进一步处理更加繁琐

hive (ods)> with tmp as (select username,age,sex,id,ids,num
          >              from jsont1
          >              lateral view json_tuple(json,'id','ids','total_number') t1 as id,ids,num)
          > select username,age,sex,id,ids1,num
          > from tmp
          > lateral view explode(split(regexp_replace(ids,"\\[|\\]",""),",")) t1 as ids1;
OK
username	age	sex	id	ids1	num
user1	18	male	1	101	3
user1	18	male	1	102	3
user1	18	male	1	103	3
user2	28	female	2	201	3
user2	28	female	2	202	3
user2	28	female	2	203	3
user3	38	male	3	301	4
user3	38	male	3	302	4
user3	38	male	3	303	4
user3	38	male	3	304	4
user4	48	female	4	401	5
user4	48	female	4	402	5
user4	48	female	4	403	5
user4	48	female	4	404	5
user4	48	female	4	405	5
user5	58	male	5	501	3
user5	58	male	5	502	3
user5	58	male	5	503	3
user6	68	female	6	601	3
user6	68	female	6	602	3
user6	68	female	6	603	3
user7	78	male	7	701	3
user7	78	male	7	702	3
user7	78	male	7	703	3
Time taken: 0.459 seconds, Fetched: 24 row(s)
hive (ods)>

(4)只能处理简单的json,复杂的json不好处理

二.使用自定义UDF函数处理json数据

1.引入依赖

<dependency>
    <groupId>org.apache.hive</groupId>
    <artifactId>hive-exec</artifactId>
    <version>2.3.7</version>
</dependency>

2.自定义UDF

public class ParseJsonArray extends UDF {
    public ArrayList<String> evaluate(String jsonStr, String key) {
        if (Strings.isNullOrEmpty(jsonStr)) {
            return null;
        }

        try {
            JSONObject jsonObject = JSON.parseObject(jsonStr);
            JSONArray jsonArray = jsonObject.getJSONArray(key);
            ArrayList<String> strings = new ArrayList<>();
            for (Object o : jsonArray) {
                strings.add(o.toString());
            }
            return  strings;
        }catch (JsonException e){
            return null;
        }
    }

    @Test
    public void junitParseJsonArray(){
        String jsonStr = "{\"id\":1,\"ids\":[101,102,103],\"total_number\":3}";
        String key = "ids";
        ArrayList<String> list = evaluate(jsonStr, key);
        System.out.println(JSON.toJSONString(list));//["101","102","103"]
    }
}

3.打包使用自定义UDF(我这里是定义临时UDF)

hive (ods)> add jar /root/data/tmp_jar/FlumeInterceptor-1.0-SNAPSHOT.jar ;
Added [/root/data/tmp_jar/FlumeInterceptor-1.0-SNAPSHOT.jar] to class path
Added resources: [/root/data/tmp_jar/FlumeInterceptor-1.0-SNAPSHOT.jar]
hive (ods)> create temporary function json_array_me as "com.lagou.dw.flume.interceptor.hive.udf.ParseJsonArray";
OK
Time taken: 2.046 seconds
hive (ods)>

4.使用自定义UDF

hive (ods)> select username,age,sex,json_array_me(json,“ids”) ids

          > from jsont1;
OK
username	age	sex	ids
user1	18	male	["101","102","103"]
user2	28	female	["201","202","203"]
user3	38	male	["301","302","303","304"]
user4	48	female	["401","402","403","404","405"]
user5	58	male	["501","502","503"]
user6	68	female	["601","602","603"]
user7	78	male	["701","702","703"]
	NULL	NULL	NULL
Time taken: 0.748 seconds, Fetched: 8 row(s)

突然发现自定义的好处:不用繁琐解决解析json字符串和普通字段不能同时显示问题.
同时我们对解析后字符数据炸裂更加方便

hive (ods)> select username,age,sex,ids1
          > from jsont1
          > lateral view explode(json_array_me(json,"ids")) t1 as ids1;
OK
username	age	sex	ids1
user1	18	male	101
user1	18	male	102
user1	18	male	103
user2	28	female	201
user2	28	female	202
user2	28	female	203
user3	38	male	301
user3	38	male	302
user3	38	male	303
user3	38	male	304
user4	48	female	401
user4	48	female	402
user4	48	female	403
user4	48	female	404
user4	48	female	405
user5	58	male	501
user5	58	male	502
user5	58	male	503
user6	68	female	601
user6	68	female	602
user6	68	female	603
user7	78	male	701
user7	78	male	702
user7	78	male	703
Time taken: 1.503 seconds, Fetched: 24 row(s)

拓展:我们如果需要将json都展开呢？
解决办法: 使用两次lateral view处理

hive (ods)> select username,age,sex,id,num,ids1
          > from jsont1
          > lateral view json_tuple(json,'id','total_number') t1 as id,num
          > lateral view explode(json_array_me(json,"ids")) t2 as ids1;
OK
username	age	sex	id	num	ids1
user1	18	male	1	3	101
user1	18	male	1	3	102
user1	18	male	1	3	103
user2	28	female	2	3	201
user2	28	female	2	3	202
user2	28	female	2	3	203
user3	38	male	3	4	301
user3	38	male	3	4	302
user3	38	male	3	4	303
user3	38	male	3	4	304
user4	48	female	4	5	401
user4	48	female	4	5	402
user4	48	female	4	5	403
user4	48	female	4	5	404
user4	48	female	4	5	405
user5	58	male	5	3	501
user5	58	male	5	3	502
user5	58	male	5	3	503
user6	68	female	6	3	601
user6	68	female	6	3	602
user6	68	female	6	3	603
user7	78	male	7	3	701
user7	78	male	7	3	702
user7	78	male	7	3	703
Time taken: 0.568 seconds, Fetched: 24 row(s)

三.对于纯json文件可以用SerDe解析

创建表可以指定表的序列化方式
TextFile:org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe
OrcSerde:org.apache.hadoop.hive.ql.io.orc.OrcSerde
Parquet:org.apache.hadoop.hive.ql.io.parquet.serde.ParquetHiveSerDe

例如:指定序列化方式org.apache.hive.hcatalog.data.JsonSerDe

[root@Linux122 source]# cat json_data.txt 
{"id":1,"ids":[101,102,103],"total_number":3}
{"id":2,"ids":[201,202,203],"total_number":3}
{"id":3,"ids":[301,302,303,304],"total_number":4}
{"id":4,"ids":[401,402,403,404,405],"total_number":5}
{"id":5,"ids":[501,502,503],"total_number":3}
{"id":6,"ids":[601,602,603],"total_number":3}
{"id":7,"ids":[701,702,703],"total_number":3}
 
hive (ods)> create table json2(
          > id int,
          > ids array<string>,
          > total_number int)
          > row format serde 'org.apache.hive.hcatalog.data.JsonSerDe';
OK
Time taken: 0.702 seconds
hive (ods)> load data local inpath '/root/data/source/json_data.txt' into table json2;
Loading data to table ods.json2
OK
Time taken: 17.651 seconds
hive (ods)> select * from json2;
OK
json2.id	json2.ids	json2.total_number
1	["101","102","103"]	3
2	["201","202","203"]	3
3	["301","302","303","304"]	4
4	["401","402","403","404","405"]	5
5	["501","502","503"]	3
6	["601","602","603"]	3
7	["701","702","703"]	3
Time taken: 0.94 seconds, Fetched: 7 row(s)
# hive # json # hadoop
Logo
华为云开发者联盟

为开发者提供学习成长、分享交流、生态实践、资源工具等服务，帮助开发者快速成长。

更多推荐

cover

解锁HDC 2024之旅：从购票到报名，全程攻略

avatar 华为云开发者联盟
cover

从原始边列表到邻接矩阵Python实现图数据处理的完整指南

avatar 华为云开发者联盟
cover

华为云云原生FinOps解决方案，释放云原生最大价值

avatar 华为云开发者联盟