JSON_EXTRACT_SCALAR возвращает нуль после разбора

Question

Я пытаюсь проанализировать JSON с массивом полей в BigQuery следующим образом:

SELECT ARRAY( SELECT STRUCT( CAST(JSON_EXTRACT_SCALAR(split_items, '$.index') AS int64) AS index, TIMESTAMP_MILLIS(CAST(JSON_EXTRACT_SCALAR(split_items, '$.startTime') AS int64)) AS startTime) FROM ( SELECT CONCAT('{', REGEXP_REPLACE(split_items, r'^"\[{|}\]"$', ''), '}') AS split_items FROM UNNEST(SPLIT(json, '},{')) AS split_items )) AS split_items FROM ( SELECT json FROM dataset:table)

Пример JSON выглядит следующим образом:

"[{\"index\":0,\"startTime\":1524607204000},{\"index\":1,\"startTime\":1524607204000},{\"index\":2,\"startTime\":1524607204000}]"

index и startTime оба равны нулю. Что-то не так с этим запросом?

Answer 1

Ниже должно работать

#standardSQL
SELECT
  ARRAY(
    SELECT 
      STRUCT( 
        CAST(JSON_EXTRACT_SCALAR(split_items, '$.index') AS int64) AS index,
          TIMESTAMP_MILLIS(CAST(JSON_EXTRACT_SCALAR(split_items, '$.startTime') AS int64)) AS startTime
      )
    FROM (
      SELECT CONCAT('{', REGEXP_REPLACE(split_items, r'\[{|}\]', ''), '}') AS split_items
      FROM UNNEST(SPLIT(json, '},{')) AS split_items 
    )
  ) AS split_items
FROM `dataset.table`     

«Исправление» находится под строкой

REGEXP_REPLACE(split_items, r'\[{|}\]', '')   

Вы можете протестировать / поиграть с ним, используя ваш пример json, как показано ниже

#standardSQL
WITH `dataset.table` AS (
  SELECT '[{"index":0,"startTime":1524607204000},{"index":1,"startTime":1524607204000},{"index":2,"startTime":1524607204000}]' AS json
)
SELECT
  ARRAY(
    SELECT 
      STRUCT( 
        CAST(JSON_EXTRACT_SCALAR(split_items, '$.index') AS int64) AS index,
          TIMESTAMP_MILLIS(CAST(JSON_EXTRACT_SCALAR(split_items, '$.startTime') AS int64)) AS startTime
      )
    FROM (
      SELECT CONCAT('{', REGEXP_REPLACE(split_items, r'\[{|}\]', ''), '}') AS split_items
      FROM UNNEST(SPLIT(json, '},{')) AS split_items 
    )
  ) AS split_items
FROM `dataset.table`

с результатом как

Row split_items.index   split_items.startTime    
1   0                   2018-04-24 22:00:04.000 UTC  
    1                   2018-04-24 22:00:04.000 UTC  
    2                   2018-04-24 22:00:04.000 UTC