SQL json_array_get - 搜索对象而不是位置

Question

我正在尝试从以下查询中的列中获取值：

SELECT 
    json_array_get(transactions.tags,0) as category,
    transactions.tags, 
    transactions.date as transaction_date,
    transactions.amount,
    CAST(transactions.amount as DECIMAL(20, 2)) as amount_banking,
    transactions.balance,
    opportunity.osc_score_previous_conduct
FROM "decisionengine"."public"."app_banking_transaction" AS "transactions"
LEFT JOIN 
    "decisionengine"."public"."app_banking_transactionbankdetail" AS "transactionsDetails" 
ON 
    "transactionsDetails".id  = "transactions".transactionBankDetail_id 
    
LEFT JOIN 
    "decisionengine"."public"."app_banking_bankingdetail" AS "banking" 
ON 
    "transactionsDetails".bankingDetail_id  = "banking".id 
    
LEFT JOIN 
    "decisionengine"."public"."app_banking_opportunitydetail" AS "opportunity" 
ON 
    "banking".opportunity_id  = "opportunity".id 
LIMIT 5

在标签列上有时会有一个带有节点类别的对象，如果该节点存在于该 JSON 中，我想提取类别的值。

我可以看到 json_array_get 允许我在数组中获得一个位置，有没有办法使用它或者我如何修改我的查询以获得类别的值？

---

编辑

WITH dataset AS (
    SELECT * FROM (SELECT tags FROM "decisionengine"."public"."app_banking_transaction" LIMIT 10) AS t (json_string)
)

SELECT
    reduce(
        CAST (json_string as ARRAY(MAP(VARCHAR, VARCHAR))),
        map(),
        (s, x) -> map_concat(s,x), 
        s -> s
    )['category']
FROM dataset

我得到了错误

Answer 1

要访问 category 值，您可以使用下一个 json 路径：$.[?(@.category)].category（将返回带有值的数组），但 Athena 中的 json 路径支持不是很好，因此您需要执行一些转换 - 首先转换 json到地图数组，然后将它们与reduce 连接到一张地图，最后从结果地图中提取category 值：

WITH dataset AS (
    SELECT * FROM (VALUES   
       ( JSON '[{"category": "test"},{"prop":2}]'),
       ( JSON '[{"id":"1"}]'),                   
       ( JSON '[{"category": "test1"}, {"prop":2}]')
 ) AS t (json_string))


SELECT
 reduce(
   CAST (json_string as ARRAY(MAP(VARCHAR, VARCHAR))),
   map(),
   (s, x) -> map_concat(s,x), 
   s -> s
 )['category']
FROM dataset

输出：

_col0
test
test1