BigQuery 和 Standard SQL：聚合数组中的每个不同字段

Question

我正在尝试计算所有五分钟时间段总和的每个地区 IP 的流量总和（以 gbps 为单位）。

我是 BigQuery 的新手，所以通过查看其他示例，我尝试了以下方法：

WITH `project.dataset.test` AS (
  SELECT '01/01/2019 12:30' time, '192.168.10.1' ip_address, 10 network, 1 gbps UNION ALL
  SELECT '01/01/2019 12:30', '192.168.10.2', 11, 2 UNION ALL
  SELECT '01/01/2019 12:30', '192.168.10.3', 12, 3 UNION ALL
  SELECT '01/01/2019 12:35', '192.168.10.1', 10, 1 UNION ALL
  SELECT '01/01/2019 12:35', '192.168.10.2', 11, 2 UNION ALL
  SELECT '01/01/2019 12:35', '192.168.10.3', 12, 3 UNION ALL
  SELECT '01/01/2019 12:40', '192.168.10.1', 10, 1 UNION ALL
  SELECT '01/01/2019 12:40', '192.168.10.2', 11, 2 UNION ALL
  SELECT '01/01/2019 12:40', '192.168.10.3', 12, 3
  ),
ip AS (
  SELECT DISTINCT (ip_address) ip_address
  FROM `project.dataset.test`
),
qualified AS (
SELECT ip_address, network, ARRAY_AGG (gbps ORDER BY ip_address DESC LIMIT 1)[SAFE_OFFSET(0)] gbps
FROM `project.dataset.test`
GROUP BY ip_address, network
)
SELECT ip_address, network, SUM(gbps)gbps
FROM (
SELECT d.ip_address ip_address, network, ARRAY_AGG (gbps ORDER BY q.ip_address DESC LIMIT 1)[SAFE_OFFSET(0)] gbps
FROM ip d
JOIN qualified q
ON q.ip_address = d.ip_address
GROUP BY ip_address, network
)
group BY ip_address, network
ORDER BY gbps DESC

我预计输出为：

Row     ip_address      network   gbps  
1       192.168.10.3    12        9
2       192.168.10.2    11        6
3       192.168.10.1    10        3

相反，实际输出是：

Row     ip_address         network   gbps   
1       192.168.10.3       12        3
2       192.168.10.2       11        2
3       192.168.10.1       10        1

我做错了什么？ select 如何计算不同 IP 的总和，而不考虑 5-5 分钟周期的数量 and/or 网络？仅供参考，我有数千行要排序，这只是我正在使用的示例。

Answer 1

How do select the sum of the distinct IPs, regardless of the number of 5-five minute periods and/or networks?

以下示例适用于 BigQuery Standatd SQL

#standardSQL
WITH `project.dataset.test` AS (
  SELECT '01/01/2019 12:30' time, '192.168.10.1' ip_address, 10 network, 1 gbps UNION ALL
  SELECT '01/01/2019 12:30', '192.168.10.2', 11, 2 UNION ALL
  SELECT '01/01/2019 12:30', '192.168.10.3', 12, 3 UNION ALL
  SELECT '01/01/2019 12:35', '192.168.10.1', 10, 1 UNION ALL
  SELECT '01/01/2019 12:35', '192.168.10.2', 11, 2 UNION ALL
  SELECT '01/01/2019 12:35', '192.168.10.3', 12, 3 UNION ALL
  SELECT '01/01/2019 12:40', '192.168.10.1', 10, 1 UNION ALL
  SELECT '01/01/2019 12:40', '192.168.10.2', 11, 2 UNION ALL
  SELECT '01/01/2019 12:40', '192.168.10.3', 12, 3
)
SELECT ip_address, network, SUM(gbps) gbps
FROM `project.dataset.test`
GROUP BY ip_address, network

结果

Row ip_address      network gbps     
1   192.168.10.3    12      9    
2   192.168.10.2    11      6    
3   192.168.10.1    10      3

BigQuery 和 Standard SQL：聚合数组中的每个不同字段

BigQuery and Standard SQL: aggregate per distnict field in array

google-bigquery

bigquery-standard-sql