hive spark 查询json中数组的长度

-- 数据格式如下
{
    "aa": "2021-09-10-08-00-00",
    "bb": "格式参考如下"
}
{
    "cc": "lpOkP...",
    "dd": [
        {
            "c1": "TGCSP...",
            "c2": 2,
            "c3": 0,
            "c4": "1436122160469676032",
            "c5": "Fw61LVxMA1BHNH5LQJH273",
            "c6": 1631233179000
        }
    ],
    "ee": [
        "1436122160469676032"
    ],
    "ff": "1436122160469676032",
    "gg": "1436122160469676032",
    "hh": 1631233179000
}

---

sparksql 添加 array 最准确,但是速度貌似不太好.还不如正则来的快. 你们自己测试下.

-- 查json中数组长度 hivesql
select
  dt,sum(size(split(regexp_replace(regexp_replace(get_json_object(json, "$.ee"),'\[|\]',''),'"',''