[嵌入式分析] 【数据准备-数据源】特殊JSON该如何解析

相信大家在使用wyn中json数据源时，觉得非常便捷。对于标准化json解析中也能得心应手的解决，但是在一些特定业务中，返回的json并不是wyn中理想的json模样。那么我们该如何去处理？

先来个练习题，看看遇到这种json我们该如何下手呢？

本来可以愉快的下班的，却遇到了这么个问题。其实json也很无辜，我也想成为你想要的样子，可是臣妾做不到啊。

其实此类json我们在教程中已经做了比较详细的解释，但是如果对jsonpath解析掌握度不够的话，那么帮助手册有==无。大家认真看以下就是教程了，干货来了：

案例1：jsonA.json 文件内容：

{"metas":{"A":{"align":"left","attrKeys":"","cname":" "}},"__sys__":{"msg":"执行成功","status":0},"data":[["A","B"],["12688","5731.00"],["8920","32667.00"],["15903","31755.00"],["16477","5228.00"],["14140","29712.00"],["13391","19130.00"],["17487","26323.00"],["22135","31505.00"],["10635","28000.00"],["24291","19201.00"],["16959","25258.00"]],"__version__":"1.0","__blocks__":{}}

操作方式：

select * from UnwindJson(@source,'$.data[1:]') tmp with(

A varchar '$.[0]',

B varchar '$.[1]'

)

解析结果：

案例2：jsonB.json 文件内容

{"__tarAppEname__":"ww","__sys__":{"msg":"执行成功","status":0},"__version__":"2.0","urlAdress":"","__blocks__":{"result":{"meta":{"columns":[{"pos":0,"name":"BRANCH_CNAME","descName":" "},{"pos":1,"name":"DATE","descName":" "},{"pos":2,"name":"POST_CODE","descName":" "},{"pos":3,"name":"PERSON_NAME","descName":" "}]},"attr":{"offset":0,"limit":10},"rows":[["工厂","202008","BB","王刚"],["工厂","202008","AA","马冲"],["工厂","202008","CC","张飞"],["工厂","202008","DD","叶开"]]}}}

操作方式：

select * from UnwindJson(@source,'$.__blocks__.result.rows

BRANCH_CNAME varchar '$.[0]',

DATE varchar '$.[1]',

POST_CODE varchar '$.[2]',

PERSON_NAME varchar '$.[3]'

)