Bigquery - json_extract数组中的所有元素

Question

我正在尝试从jsons数组中的每个json中提取两个键（使用sql legacy），目前我正在使用json extract函数：

json_extract(json_column , '$[1].X') AS X,
json_extract(json_column , '$[1].Y') AS Y,

我如何让它在'json数组列'的每个json上运行，而不仅仅是[1]（例如）？

一个例子json：

[

{"blabla":000,"X":1,"blabla":000,"blabla":000,"blabla":000,,"Y":"2"},

{"blabla":000,"X":3,"blabla":000,"blabla":000,"blabla":000,,"Y":"4"},

]   

提前致谢！

Answer 1

让我们从一个类似的问题开始 - 这不是一个从json数组中提取所有电子邮件的非常方便的方法：

SELECT id
  , [ JSON_EXTRACT_SCALAR(JSON_EXTRACT(payload, '$.commits'), '$[0].author.email')  
      , JSON_EXTRACT_SCALAR(JSON_EXTRACT(payload, '$.commits'), '$[1].author.email')  
      , JSON_EXTRACT_SCALAR(JSON_EXTRACT(payload, '$.commits'), '$[2].author.email')  
      , JSON_EXTRACT_SCALAR(JSON_EXTRACT(payload, '$.commits'), '$[3].author.email')
    ] emails
FROM `githubarchive.day.20180830` 
WHERE type='PushEvent' 
AND id='8188163772'

我们现在处理这个问题的最好方法是在UDF中使用一些JavaScript将json数组拆分成SQL数组：

CREATE TEMP FUNCTION json2array(json STRING)
RETURNS ARRAY<STRING>
LANGUAGE js AS """
  return JSON.parse(json).map(x=>JSON.stringify(x));
"""; 

SELECT * EXCEPT(array_commits),
  ARRAY(SELECT JSON_EXTRACT_SCALAR(x, '$.author.email') FROM UNNEST(array_commits) x) emails
FROM (
  SELECT id
    , json2array(JSON_EXTRACT(payload, '$.commits')) array_commits
  FROM `githubarchive.day.20180830` 
  WHERE type='PushEvent' 
  AND id='8188163772'
)

Answer 2

下面是BigQuery Standard SQL的示例，允许您接近使用JSONPath的标准方式，无需额外操作，因此您只需使用CUSTOM_JSON_EXTRACT(json, json_path)函数

#standardSQL
CREATE TEMPORARY FUNCTION CUSTOM_JSON_EXTRACT(json STRING, json_path STRING)
RETURNS ARRAY<STRING>
LANGUAGE js AS """
        return jsonPath(JSON.parse(json), json_path);
"""
OPTIONS (
    library="gs://your_bucket/jsonpath-0.8.0.js"
);
WITH t AS (
SELECT '''
[
{"blabla1":1,"X":1,"blabla2":3,"blabla3":5,"blabla4":7,"Y":"2"},
{"blabla1":2,"X":3,"blabla2":4,"blabla3":6,"blabla4":8,"Y":"4"}
]   
''' AS json_column 
)
SELECT 
  CUSTOM_JSON_EXTRACT(json_column , '$[*].X') AS X,
  CUSTOM_JSON_EXTRACT(json_column , '$[*].Y') AS Y
FROM t   

结果将是

Row X   Y    
1   1   2    
    3   4      

注意：要克服当前BigQuery对JsonPath的“限制”，上面的解决方案使用custom function和external library - jsonpath-0.8.0.js，可以从https://code.google.com/archive/p/jsonpath/downloads下载并上传到Google Cloud Storage - gs：//your_bucket/jsonpath-0.8.0.js

重新阅读Felipe的答案 - 上面他的例子解决方案将如下所示（就像FYI一样）

SELECT 
  id, 
  CUSTOM_JSON_EXTRACT(payload, '$.commits[*].author.email') emails
FROM `githubarchive.day.20180830` 
WHERE type='PushEvent' 
AND id='8188163772'