Question

我有一个只有一列（raw）的Athena表（json）。

我有以下查询，它输出json键的频率：

SELECT key, count(*)
FROM (
  SELECT map_keys(cast(json_parse(json) AS map(varchar, json))) AS keys
  FROM raw
)
CROSS JOIN UNNEST (keys) AS t (key)
GROUP BY key

如何扩展此查询，以便告诉我特定键是否具有带有任何非数字字符的值？

[我在找到答案后删除了失败的尝试]

Answer 1

这有效：

SELECT k, count(*) as isPresent, sum(isNumber) as isNumber, 
          count(*)-sum(isNumber) as notIsNumber from (
               with dataset as (SELECT 
                   cast(json_parse(json) AS map(varchar, varchar)) as kv FROM raw)
               SELECT t.k, t.v, 
                      IF(TRY(cast(t.v as double)) is null, 0, 1) as isNumber 
               from dataset cross join unnest(kv) as t(k, v)
) group by k

Athena / Presto数据发现查询可推荐JSON模式？

1 个答案: