如何通过Rails 3中的时间片(总和,平均值,最小值,最大值等)获取聚合数据

时间:2022-02-03 20:57:13

How can I create active relation queries in Rails 3 that are aggregated by time slice?

如何在Rails 3中创建由时间片聚合的活动关系查询?

I'd like to build queries that for every n minute interval can return min, max, avg, sum, count, for every sample of a counter with a particular name.

我想建立一个查询,每隔n分钟可以为具有特定名称的计数器的每个样本返回min,max,avg,sum,count。

create_table "samples"  
    t.integer  "counter_id"  
    t.string "name"  
    t.float    "value"  
    t.datetime "created_at"  
end  

2 个解决方案

#1


4  

Unfortunately I've never used Postgres, so this solution works in MySQL. But I think you can find out Postgres analogs.

不幸的是我从未使用过Postgres,所以这个解决方案适用于MySQL。但我认为你可以找到Postgres类似物。

class Counter < ActiveRecord::Base
  has_many :samples do
    # default 30 minutes
    def per_time_slice(slice = 30)
      start = "2000-01-01 00:00:00"
      self.select("*, 
                   CONCAT( FLOOR(TIMESTAMPDIFF(MINUTE,'#{start}',created_at)/#{slice})*#{slice},  
                     (FLOOR(TIMESTAMPDIFF(MINUTE,'#{start}',created_at)/#{slice})+1)*#{slice} ) as slice,
                   avg(value) as avg_value, 
                   min(value) as min_value, 
                   max(value) as max_value, 
                   sum(value) as sum_value, 
                   count(value) as count_value").
                   group("slice").order("slice")
    end
  end
end

Usage

counter = find_some_counter
samples = counter.samples.per_time_slice(60).where(:name => "Bobby")
samples.map(&:avg_value)
samples.map(&:min_value)
samples.map(&:max_value)

etc

#2


1  

This should do the trick:

这应该是诀窍:

Samples.where("SQL for the time slice goes here").where(:name, "Views").average(:value)

You can also swap average for any of minimum, maximum and sum.

您还可以将平均值换成最小值,最大值和总和。

#1


4  

Unfortunately I've never used Postgres, so this solution works in MySQL. But I think you can find out Postgres analogs.

不幸的是我从未使用过Postgres,所以这个解决方案适用于MySQL。但我认为你可以找到Postgres类似物。

class Counter < ActiveRecord::Base
  has_many :samples do
    # default 30 minutes
    def per_time_slice(slice = 30)
      start = "2000-01-01 00:00:00"
      self.select("*, 
                   CONCAT( FLOOR(TIMESTAMPDIFF(MINUTE,'#{start}',created_at)/#{slice})*#{slice},  
                     (FLOOR(TIMESTAMPDIFF(MINUTE,'#{start}',created_at)/#{slice})+1)*#{slice} ) as slice,
                   avg(value) as avg_value, 
                   min(value) as min_value, 
                   max(value) as max_value, 
                   sum(value) as sum_value, 
                   count(value) as count_value").
                   group("slice").order("slice")
    end
  end
end

Usage

counter = find_some_counter
samples = counter.samples.per_time_slice(60).where(:name => "Bobby")
samples.map(&:avg_value)
samples.map(&:min_value)
samples.map(&:max_value)

etc

#2


1  

This should do the trick:

这应该是诀窍:

Samples.where("SQL for the time slice goes here").where(:name, "Views").average(:value)

You can also swap average for any of minimum, maximum and sum.

您还可以将平均值换成最小值,最大值和总和。