多COUNT（DISTINCT）作为GROUP BY在Rails的

Question

我试图运行此查询或东西，将产生从Rails的输出相似：

SELECT 
COUNT(DISTINCT columnName = 'Value1') AS Value1, 
COUNT(DISTINCT columnName = 'Value2') AS Value2, 
COUNT(DISTINCT columnName = 'Value3') AS Value3, 
COUNT(DISTINCT columnName = 'Value4') AS Value4, 
COUNT(DISTINCT columnName = 'Value5') AS Value5, 
COUNT(DISTINCT columnName = 'Value6') AS Value6, 
COUNT(DISTINCT columnName = 'Value7') AS Value7, 
COUNT(DISTINCT columnName = 'Value8') AS Value8, 
COUNT(DISTINCT columnName = 'Value9') AS Value9, 
COUNT(DISTINCT columnName = 'Value10') AS Value10, 
COUNT(DISTINCT columnName = 'Value11') AS Value11, 
COUNT(DISTINCT columnName = 'Value12') AS Value12, 
COUNT(DISTINCT columnName = 'Value13') AS Value13, 
COUNT(DISTINCT columnName = 'Value14') AS Value14, 
COUNT(DISTINCT columnName = 'Value15') AS Value15, 
unix_timestamp(timestamp) div 300 AS unix_timestamp_timestamp_div_300 
FROM `table` WHERE (timestamp > '2013-12-15 15:50:00 -0800') 
GROUP BY unix_timestamp(timestamp) div 300 

我已经试过.Count中的多个版本的选项，并没有和.group，都没有运气。

编辑： 为了澄清，我们的目标是每5分钟获得一行，并计算5分钟内每个Value＃标签有多少行。这将被用来为每个Value＃创建一个包含数据序列的图。

编辑2戈登Linhoff指出的下方，查询应该是：

SELECT 
    sum(columnName = 'Value1') AS Value1, 
    sum(columnName = 'Value2') AS Value2, 
    sum(columnName = 'Value3') AS Value3, 
    sum(columnName = 'Value4') AS Value4, 
    sum(columnName = 'Value5') AS Value5, 
    sum(columnName = 'Value6') AS Value6, 
    sum(columnName = 'Value7') AS Value7, 
    sum(columnName = 'Value8') AS Value8, 
    sum(columnName = 'Value9') AS Value9, 
    sum(columnName = 'Value10') AS Value10, 
    sum(columnName = 'Value11') AS Value11, 
    sum(columnName = 'Value12') AS Value12, 
    sum(columnName = 'Value13') AS Value13, 
    sum(columnName = 'Value14') AS Value14, 
    sum(columnName = 'Value15') AS Value15, 
    unix_timestamp(timestamp) div 300 AS unix_timestamp_timestamp_div_300 
FROM `table` 
WHERE (timestamp > '2013-12-15 15:50:00 -0800') 
GROUP BY unix_timestamp(timestamp) div 300 

因此，MySQL是现在固定，在获取的Rails返回的东西，除了一堆=>零个对象模型的任何帮助吗？

Answer 1

我想出了内在戈登的回答后，下面的解决方案。

pollEvery = 300 # Five Minutes 
graphLength = 2592000 # About a month 
now = (Time.now.utc.to_i/pollEvery) * pollEvery 
start_time = now - graphLength 

distinctValues = ModelName.where("created_at > ?", start_time).uniq.pluck(:columnName) 

sumClauses = Array.new 
distinctValues.each do |value| 
    clause = "SUM(hangup_cause = '#{value}') AS #{value}" 
    sumClauses.push(clause) 
end 

selectClause = sumClauses.join(', ') + ", created_at" 

data = ModelName.group("unix_timestamp(created_at) div #{pollEvery}").select(selectClause).where('created_at > ?', Time.at(start_time)) 

data.each do |row| 
    # You can loop through the distinctValues to get the data 
    # using row[value] in a distinctValues.each do |value| loop 
    # The created_at timestamp for each grouping is available as well 
end 

Answer 2

您的count(distinct)将返回0或1，具体取决于行是否与条件匹配。也就是说，只有条件返回两个值，0或1更改count(distinct)到sum()：

SELECT 
    sum(columnName = 'Value1') AS Value1, 
    sum(columnName = 'Value2') AS Value2, 
    sum(columnName = 'Value3') AS Value3, 
    sum(columnName = 'Value4') AS Value4, 
    sum(columnName = 'Value5') AS Value5, 
    sum(columnName = 'Value6') AS Value6, 
    sum(columnName = 'Value7') AS Value7, 
    sum(columnName = 'Value8') AS Value8, 
    sum(columnName = 'Value9') AS Value9, 
    sum(columnName = 'Value10') AS Value10, 
    sum(columnName = 'Value11') AS Value11, 
    sum(columnName = 'Value12') AS Value12, 
    sum(columnName = 'Value13') AS Value13, 
    sum(columnName = 'Value14') AS Value14, 
    sum(columnName = 'Value15') AS Value15, 
    unix_timestamp(timestamp) div 300 AS unix_timestamp_timestamp_div_300 
FROM `table` 
WHERE (timestamp > '2013-12-15 15:50:00 -0800') 
GROUP BY unix_timestamp(timestamp) div 300