避免 BigQuery 中的相关子查询错误

Question

我有一个简单的查询来获取创建交易时使用的货币汇率：

SELECT t.orderid, t.date, 
 (SELECT rate FROM sources.currency_rates r WHERE currencyid=1 AND 
r.date>=t.date  ORDER BY date LIMIT 1) rate
FROM sources.transactions t

这会触发错误：

Error: Correlated subqueries that reference other tables are not 
supported unless they can be de-correlated, such as by transforming 
them into an efficient JOIN.' 

我尝试了几种类型的连接和命名子查询，但似乎都没有。实现这一目标的最佳方法是什么？似乎是一个非常常见的场景，应该很容易在 BQ 的标准 Sql 中实现。

Answer 1

以下是 BigQuery 标准 SQL

#standardSQL
SELECT 
  t.orderid AS orderid, 
  t.date AS date, 
  ARRAY_AGG(r.rate ORDER BY r.date LIMIT 1)[SAFE_OFFSET(0)] AS rate
FROM `sources.transactions` AS t
JOIN `sources.currency_rates` AS r
ON currencyid = 1 
AND r.date >= t.date
GROUP BY orderid, date

Answer 2

我注意到其他相关子查询的类似行为。它们很有用，但并不总是能被 BigQuery 自动建模为 JOIN。

类似的情况：

#standardSQL
SELECT name, (
  SELECT AVG(temp) 
  FROM `bigquery-public-data.noaa_gsod.gsod2017` b
  WHERE a.usaf=b.stn  
) temp
FROM `bigquery-public-data.noaa_gsod.stations` a
LIMIT 10

不起作用：

#standardSQL
SELECT name, (
  SELECT temp 
  FROM `bigquery-public-data.noaa_gsod.gsod2017` b
  WHERE a.usaf=b.stn  
  ORDER BY da 
  LIMIT 1
) temp
FROM `bigquery-public-data.noaa_gsod.stations` a
LIMIT 10

修复：

#standardSQL
SELECT name, ARRAY_AGG(temp ORDER BY da LIMIT 1) temp
FROM `bigquery-public-data.noaa_gsod.stations` a
JOIN `bigquery-public-data.noaa_gsod.gsod2017` b
ON a.usaf=b.stn  
GROUP BY 1
LIMIT 10

（给我一个公共数据集，我会编写一个适用于您的数据的查询）