hive之Json解析(普通Json和Json数组)
一、数据准备
现准备原始json数据(test.json)如下：
{"movie":"1193","rate":"5","timeStamp":"978300760","uid":"1"}
{"movie":"661","rate":"3","timeStamp":"978302109","uid":"1"}
{"movie":"914","rate":"3","timeStamp":"978301968","uid":"1"}
{"movie":"3408","rate":"4","timeStamp":"978300275","uid":"1"}
{"movie":"2355","rate":"5","timeStamp":"978824291","uid":"1"}
{"movie":"1197","rate":"3","timeStamp":"978302268","uid":"1"}
{"movie":"1287","rate":"5","timeStamp":"978302039","uid":"1"}
{"movie":"2804","rate":"5","timeStamp":"978300719","uid":"1"}
{"movie":"594","rate":"4","timeStamp":"978302268","uid":"1"}
现在将数据导入到hive中，并且最终想要得到这么一个结果：
可以使用：内置函数（get_json_object）或者自定义函数完成
二、get_json_object(string json_string, string path)
返回值：String
说明：解析json的字符串json_string，返回path指定的内容。如果输入的json字符串无效，那么返回NUll，这个函数每次只能返回一个数据项。
0: jdbc:hive2://hadoop3:10000> select get_json_object(\'{"movie":"594","rate":"4","timeStamp":"978302268","uid":"1"}\',\'$.movie\');
1、创建json表并将数据导入
0: jdbc:hive2://master:10000> create table json(data string); No rows affected (0.572 seconds)
0: jdbc:hive2://master:10000> load data local inpath \'/home/hadoop/json.txt\' into table json;
No rows affected (1.046 seconds)
0: jdbc:hive2://master:10000> select get_json_object(data,\'$.movie\') as movie from json;
三、json_tuple(jsonStr, k1, k2, …)
参数为一组键k1，k2，。。。。。和json字符串，返回值的元组。该方法比get_json_object高效，因此可以在一次调用中输入多次键
0: jdbc:hive2://master:10000> select b.b_movie,b.b_rate,b.b_timeStamp,b.b_uid from json a lateral view
json_tuple(a.data,\'movie\',\'rate\',\'timeStamp\',\'uid\') b as b_movie,b_rate,b_timeStamp,b_uid;
注意点：
json_tuple相当于get_json_object的优势就是一次可以解析多个Json字段。但是如果我们有个Json数组，这两个函数都无法处理
四、Json数组解析
1、使用Hive自带的函数解析Json数组
Hive的内置的explode函数，explode()函数接收一个 array或者map 类型的数据作为输入，然后将 array 或 map 里面的元素按照每行的形式输出。其可以配合 LATERAL VIEW 一起使用。
hive> select explode(array(\'A\',\'B\',\'C\')); OK A B C Time taken: 4.879 seconds, Fetched: 3 row(s) hive> select explode(map(\'A\',10,\'B\',20,\'C\',30)); OK A 10 B 20 C 30 Time taken: 0.261 seconds, Fetched: 3 row(s)
这个explode函数和我们解析json数据是有关系的，我们可以使用explode函数将json数组里面的元素按照一行一行的形式输出：
hive> SELECT explode(split(regexp_replace(regexp_replace(\'[{"website":"www.baidu.com","name":"百度"},{"website":"google.com","name":"谷歌"}]\', \'\\]\',\'\'),\'\\}\\,\\{\',\'\\}\\;\\{\'),\'\\;\')); OK {"website":"www.baidu.com","name":"百度"} {"website":"google.com","name":"谷歌"} Time taken: 0.14 seconds, Fetched: 2 row(s)
说明:
SELECT explode(split( regexp_replace( regexp_replace( \'[ {"website":"www.baidu.com","name":"百度"}, {"website":"google.com","name":"谷歌"} ]\', \'\\[|\\]\',\'\'), --将 Json 数组两边的中括号去掉 \'\\}\\,\\{\' --将 Json 数组元素之间的逗号换成分号 ,\'\\}\\;\\{\'), \'\\;\')); --以分号作为分隔符
结合 get_json_object 或 json_tuple 来解析里面的字段：
hive> select json_tuple(json, \'website\', \'name\') from (SELECT explode(split(regexp_replace(regexp_replace(\'[{"website":"www.baidu.com","name":"百},{"website":"google.com","name":"谷歌"}]\', \'\\[|\\]\',\'\'),\'\\}\\,\\{\',\'\\}\\;\\{\'),\'\\;\')) as json) test; OK www.baidu.com 百度 google.com 谷歌 Time taken: 0.283 seconds, Fetched: 2 row(s)
2、自定义函数解析JSON数组
虽然可以使用Hive自带的函数类解析Json数组，但是使用起来有些麻烦。Hive提供了强大的自定义函数（UDF）的接口，我们可以使用这个功能来编写解析JSON数组的UDF。具体测试过程如下：
<dependencies>
<dependency>
<groupId>org.apache.hive</groupId>
<artifactId>hive-exec</artifactId>
<version>2.1.1</version>
</dependency>
</dependencies>
import org.apache.hadoop.hive.ql.exec.Description; import org.apache.hadoop.hive.ql.exec.UDF; import org.json.JSONArray; import org.json.JSONException; import java.util.ArrayList; @Description(name = "json_array", value = "_FUNC_(array_string) - Convert a string of a JSON-encoded array to a Hive array of strings.") public class JsonArray extends UDF{ public ArrayList<String> evaluate(String jsonString) { if (jsonString == null) { return null; } try { JSONArray extractObject = new JSONArray(jsonString); ArrayList<String> result = new ArrayList<String>(); for (int ii = 0; ii < extractObject.length(); ++ii) { result.add(extractObject.get(ii).toString()); } return result; } catch (JSONException e) { return null; } catch (NumberFormatException e) { return null; } } }
将上面的代码进行编译打包,jar包名为：HiveJsonTest-1.0-SNAPSHOT.jar
hive> add jar /mnt/HiveJsonTest-1.0-SNAPSHOT.jar; Added [/mnt/HiveJsonTest-1.0-SNAPSHOT.jar] to class path Added resources: [/mnt/HiveJsonTest-1.0-SNAPSHOT.jar]
hive> create temporary function json_array as \'JsonArray\'; OK Time taken: 0.111 seconds
hive> select explode(json_array(\'[{"website":"www.baidu.com","name":"百度"},{"website":"google.com"name":"谷歌"}]\')); OK {"website":"www.baidu.com","name":"百度"} {"website":"google.com","name":"谷歌"} Time taken: 10.427 seconds, Fetched: 2 row(s)
hive> select json_tuple(json, \'website\', \'name\') from (SELECT explode(json_array(\'[{"website":"www.baidu.com","name":"百度"},{"website":"google.com","name":"谷歌"}]\')) as json) test; OK www.baidu.com 百度 google.com 谷歌 Time taken: 0.265 seconds, Fetched: 2 row(s)
3、自定义函数解析json对象
package com.laotou; import org.apache.commons.lang3.StringUtils; import org.apache.hadoop.hive.ql.exec.UDF; import org.json.JSONException; import org.json.JSONObject; import org.json.JSONTokener; /** * * add jar jar/bdp_udf_demo-1.0.0.jar; * create temporary function getJsonObject as \'com.laotou.JsonObjectParsing\'; * Json对象解析UDF * @Author: * @Date: 2019/8/9 */ public class JsonObjectParsing extends UDF { public static String evaluate(String jsonStr, String keyName) throws JSONException { if(StringUtils.isBlank(jsonStr) || StringUtils.isBlank(keyName)){ return null; } JSONObject jsonObject = new JSONObject(new JSONTokener(jsonStr)); Object objValue = jsonObject.get(keyName); if(objValue==null){ return null; } return objValue.toString(); } }
3、1准备数据
3、2测试
版权声明：本文为yfb918原创文章，遵循 CC 4.0 BY-SA 版权协议，转载请附上原文出处链接和本声明。