无法解析 Google BigQuery 中的日期信息

Question

我已将一些数据从 MongoDB 移动到 BigQuery。不幸的是，现在无法访问 MongoDB 源。我能够在stackoverflow community、Alessandro 的帮助下解析大部分信息，但我在解析日期信息方面遇到了问题。以下是我创建的示例数据

with cte as (
    select 12829 as user_id,
    '[{"_id":{"$oid":"5d650676af82eb0a30737e74"},"_type":"UserName","capture_date":{"$date":"2019-08-27T00:00:00.000+0000"},"source":"google","name1":"John","name3":"Doe","name4":"D/O Jane Doe","gender":"1","dob":{"$date":"1986-07-10T00:00:00.000+0000"}}]' as json_line
)
select
    user_id,
    json_value(json_line, '$[0].name1') as name1,
    json_value(json_line, '$[0].capture_date') as capture_date,
    PARSE_DATE('{"$date":"%Y-%m-%dT00:00:00.000+0000"}', json_value(json_line, '$[0].capture_date')) as capture_date1
from cte

我已经尝试过使用 parse_date，它没有用。您能否在提供的方法中帮助我解决我的错，或者是否有任何更好的方法来解决同样的问题？

我正在尝试解析 cature_date 信息并且得到空值。

Answer 1

试试json_extract_scalar:

with cte as (
    select 12829 as user_id,
    '[{"_id":{"$oid":"5d650676af82eb0a30737e74"},"_type":"UserName","capture_date":{"$date":"2019-08-27T00:00:00.000+0000"},"source":"google","name1":"John","name3":"Doe","name4":"D/O Jane Doe","gender":"1","dob":{"$date":"1986-07-10T00:00:00.000+0000"}}]' as json_line
)
select
    user_id,
    json_extract(json_line, '$[0].name1') as name1,
    json_extract_scalar(json_line, "$[0].capture_date['$date']") as capture_date,
    PARSE_DATE("%Y-%m-%dT00:00:00.000+0000", json_extract_scalar(json_line, "$[0].capture_date['$date']")) as capture_date1,
from cte

或json_value:

with cte as (
    select 12829 as user_id,
    '[{"_id":{"$oid":"5d650676af82eb0a30737e74"},"_type":"UserName","capture_date":{"$date":"2019-08-27T00:00:00.000+0000"},"source":"google","name1":"John","name3":"Doe","name4":"D/O Jane Doe","gender":"1","dob":{"$date":"1986-07-10T00:00:00.000+0000"}}]' as json_line
)
select
    user_id,
    json_extract(json_line, '$[0].name1') as name1,
    json_value(json_line, '$[0].capture_date."$date"') as capture_date,
    PARSE_DATE("%Y-%m-%dT00:00:00.000+0000", json_value(json_line, '$[0].capture_date."$date"')) as capture_date1,
from cte