1

说我查询表

rawStats = Stats::TrackProcessed.select("token")

并收到表中所有令牌的数量。

看起来像这样：

3e79a387c29bda1069271e06ad03d82b8296242e
059681f46ab1c1fa8cf8443a82f0898172e0b646
eacd846ea4e91b49f92f416f61e0f2d075b9dae7
eacd846ea4e91b49f92f416f61e0f2d075b9dae7
811705019a970929801adbf3db0ede31ed01816c

我需要返回一个看起来像这样的哈希列表

{
   '3e79a387c29bda1069271e06ad03d82b8296242e' => 1,
   '059681f46ab1c1fa8cf8443a82f0898172e0b646' => 1
   'eacd846ea4e91b49f92f416f61e0f2d075b9dae7' => 2
   '811705019a970929801adbf3db0ede31ed01816c' => 1
}

其中第一个值是一个标记，第二个值是该标记在表中的出现次数。然后我需要从所有记录中得到这个平均值。

到目前为止，我已经得到了查询表

rawStats = Stats::TrackProcessed.select("distinct token")

通过检索唯一令牌

然后虽然每个循环计算每个令牌的出现次数

rawArr = []
rawStats.each do |r|
  token = {
    :token => r.token,
    :count => rawStats.where("token = ?",r.token).count('token')
  }

  rawArr << token
end

但这过程为

(0.7ms)  SELECT COUNT(`stats_track_processed`.`token`) FROM `stats_track_processed` WHERE `stats_track_processed`.`token` = '196e595b573f71fc2af04693c73809303bebd62d'
(0.7ms)  SELECT COUNT(`stats_track_processed`.`token`) FROM `stats_track_processed` WHERE `stats_track_processed`.`token` = 'db67ab44e94ca338d90e902a36c37b4998a47ff0'
(0.7ms)  SELECT COUNT(`stats_track_processed`.`token`) FROM `stats_track_processed` WHERE `stats_track_processed`.`token` = 'a8a78cffc0935b07f90b5f3008dcad27e8ac71c7'
(0.8ms)  SELECT COUNT(`stats_track_processed`.`token`) FROM `stats_track_processed` WHERE `stats_track_processed`.`token` = '91b8ae5ffa12ece30d8548b488fab8aff7614f2b'
(0.7ms)  SELECT COUNT(`stats_track_processed`.`token`) FROM `stats_track_processed` WHERE `stats_track_processed`.`token` = '649e7aa59cc8af6c59cec1cf637ffed1ce6b3be7'
(0.7ms)  SELECT COUNT(`stats_track_processed`.`token`) FROM `stats_track_processed` WHERE `stats_track_processed`.`token` = '441dae6f22776687b57daaaeef2c63c31902b987'
(0.7ms)  SELECT COUNT(`stats_track_processed`.`token`) FROM `stats_track_processed` WHERE `stats_track_processed`.`token` = '6b8937a13012f22c8a4cdd4ed4caad0ad8761d3b'
(0.6ms)  SELECT COUNT(`stats_track_processed`.`token`) FROM `stats_track_processed` WHERE `stats_track_processed`.`token` = 'beaba154f38d5c9b64e4fb2d851ec785bd6bc4ec'
(0.6ms)  SELECT COUNT(`stats_track_processed`.`token`) FROM `stats_track_processed` WHERE `stats_track_processed`.`token` = 'efa9f3a71c4ff33abcae4af788ef6d0be599f76a'
(0.7ms)  SELECT COUNT(`stats_track_processed`.`token`) FROM `stats_track_processed` WHERE `stats_track_processed`.`token` = '053a7727885f3c4099a9b31a2eccc77b99df50ee'
(0.7ms)  SELECT COUNT(`stats_track_processed`.`token`) FROM `stats_track_processed` WHERE `stats_track_processed`.`token` = '32dc984c9f85ff91242533144cd179a9b4529bed'
(0.7ms)  SELECT COUNT(`stats_track_processed`.`token`) FROM `stats_track_processed` WHERE `stats_track_processed`.`token` = 'a58804ae38490f08a1645c8271e88df3757fe771f'
(0.7ms)  SELECT COUNT(`stats_track_processed`.`token`) FROM `stats_track_processed` WHERE `stats_track_processed`.`token` = 'a58804a2e38490f08a1645c8271e88df3757fe771f'

通过在处理大量数据时分别查询每个令牌，这看起来确实是一个非常糟糕的主意。

我已经通过 SQL 查询完成了这个，看起来像这样

SELECT AVG(viewsAverage) as total 
FROM 
 (SELECT COUNT(token) AS viewsAverage 
  FROM stats_track_processed 
  WHERE admedia_id = #{params[:admedia_id]} 
    AND #{params[:banner_id]} 
    AND (access_time BETWEEN '#{params[:begin]}' 
    AND DATE_ADD('#{params[:end]}', INTERVAL 1 DAY)) 
  GROUP BY token) 
stats_track_processed

但这看起来确实是一个糟糕的解决方案，我目前正在重构代码并将所有逻辑移至控制器。

任何帮助都感激不尽。

先感谢您。

4

3 回答 3

2

试试这个

 Stats::TrackProcessed.select(:token).count(group: :token)

干杯。

于 2012-08-09T07:44:52.520 回答
1

您可以在单个查询中实现此目的：

SELECT SUM(IF(token = '196e595b573f71fc2af04693c73809303bebd62d', 1, 0)) AS token_1,
       SUM(IF(token = 'db67ab44e94ca338d90e902a36c37b4998a47ff0', 1, 0)) AS token_2,
       ...
FROM `stats_track_processed`
WHERE `stats_track_processed`.`token` IN('196e595b573f71fc2af04693c73809303bebd62d',
                                         'db67ab44e94ca338d90e902a36c37b4998a47ff0', ...);

或者

SELECT token, 
       COUNT(token) AS token_count
FROM stats_track_processed
GROUP BY token;
于 2012-08-09T07:20:08.943 回答
1

遍历完整列表并维护计数的散列。

result = Stats::TrackProcessed.select("token")
counts = Hash.new{|h,k| h[k] = 0 }
result.each{|el| counts[el] +=1}

然后为了得到一个平均值，你可以做类似的事情

 average = counts.values.inject(0){|sum, el| sum + el} / counts.length
于 2012-08-09T07:28:54.087 回答