Postgres按日期（时间戳）和总和值分组。

Question

我有以下表格条目。

mydb=> SELECT username, input_bytes, output_bytes, updated_at from radius_acct;

 username | input_bytes | output_bytes |       updated_at
----------+-------------+--------------+------------------------
 sandeep  |      255081 |        42154 | 2014-05-08 09:34:13+00
 sandeep  |      250281 |        42254 | 2014-05-09 09:34:14+00
 prashant |      258081 |        42354 | 2014-05-10 09:34:15+00
 arun     |         493 |         1215 | 2014-05-17 11:16:53+00
 arun     |        9421 |         3657 | 2014-05-17 11:21:38+00
 arun     |       18325 |         4504 | 2014-05-17 17:12:53+00
 prashant |      254081 |        42454 | 2014-05-11 09:34:16+00
 raj      |      255081 |        42554 | 2014-05-12 09:34:17+00
 raj      |      250681 |        42554 | 2014-05-13 09:34:18+00
 raj      |      253081 |        42654 | 2014-05-14 09:34:19+00
 kumar    |      295081 |        47254 | 2014-05-15 09:34:16+00
(11 rows)

我想将输入和输出字节相加并以MB显示，并按日期对用户名进行分组。

我的查询不对日期进行分组并对其求和，并显示类似的输出，如下所示：

mydb=> SELECT username, input_bytes, output_bytes, ROUND((SUM(input_bytes + output_bytes)/1048576)::decimal, 2) AS total, date(updated_at)  from radius_acct group by username, input_bytes, output_bytes, updated_at ORDER BY username ASC ;
 username | input_bytes | output_bytes | total |    date
----------+-------------+--------------+-------+------------
 arun     |         493 |         1215 |  0.00 | 2014-05-17
 arun     |        9421 |         3657 |  0.01 | 2014-05-17
 arun     |       18325 |         4504 |  0.02 | 2014-05-17
 kumar    |      295081 |        47254 |  0.33 | 2014-05-15
 prashant |      254081 |        42454 |  0.28 | 2014-05-11
 prashant |      258081 |        42354 |  0.29 | 2014-05-10
 raj      |      250681 |        42554 |  0.28 | 2014-05-13
 raj      |      253081 |        42654 |  0.28 | 2014-05-14
 raj      |      255081 |        42554 |  0.28 | 2014-05-12
 sandeep  |      250281 |        42254 |  0.28 | 2014-05-09
 sandeep  |      255081 |        42154 |  0.28 | 2014-05-08

预期输出：

username | input_bytes | output_bytes | total |    date
----------+-------------+--------------+-------+------------
 **arun     |       28239 |         9376 |  0.03 | 2014-05-17**
 kumar    |      295081 |        47254 |  0.33 | 2014-05-15
 prashant |      254081 |        42454 |  0.28 | 2014-05-11
 prashant |      258081 |        42354 |  0.29 | 2014-05-10
 raj      |      250681 |        42554 |  0.28 | 2014-05-13
 raj      |      253081 |        42654 |  0.28 | 2014-05-14
 raj      |      255081 |        42554 |  0.28 | 2014-05-12
 sandeep  |      250281 |        42254 |  0.28 | 2014-05-09
 sandeep  |      255081 |        42154 |  0.28 | 2014-05-08

第一行arun总共三行。

Answer 1

尝试一次

SELECT username, 
sum(input_bytes) as totalinputbytes, 
sum(output_bytes) as totaloutputbytes, 
ROUND((SUM(input_bytes + output_bytes)/1048576)::decimal, 2) AS total, 
date(updated_at) as update_date
from radius_acct 
group by username, date(updated_at)
ORDER BY username;