实施例

Question

我有一张桌子，我会计算一个元素出现的频率。为此，我可以使用Sequel::Dataset#group_and_count。

但现在我想算一下这个结果。我怎么能这样做？

实施例

require 'sequel'
Sequel.extension :pretty_table
DB = Sequel.sqlite()

DB.create_table( :test ){
  add_field :key,    :type => :nvarchar, :size => 10
  add_field :value, :type => :nvarchar, :size => 10
}

'a'.upto('f'){|a|
  a.upto('x'){|b|
    DB[:test].insert( :key => a, :value => b)
  }
}

sel = DB[:test].group_and_count(:key)
puts  Sequel::PrettyTable.string(sel)

这给了我：

+-----+---+
|count|key|
+-----+---+
|   24|a  |
|   23|b  |
|   22|c  |
|   21|d  |
|   20|e  |
|   19|f  |
+-----+---+

现在我想计算一下计数值的发生频率。

我的结果应该是：

+-----------+-----+
|count_value|count|
+-----------+-----+
|   24      |  1  |
|   23      |  1  |
|   22      |  1  |
|   21      |  1  |
|   20      |  1  |
|   19      |  1  |
+-----------+-----+

我可以通过以下方式获得此结果：

mycount = Hash.new(0)
DB[:test].group_and_count(:key).each{| res |
  mycount[res[:count]] += 1
}
p mycount # {24=>1, 23=>1, 22=>1, 21=>1, 20=>1, 19=>1}

使用SQL（参见SQL Group by Count of Counts）我可以执行以下命令：

puts DB.fetch(<<-sql ).all
  select count_value, count(*) as count from (
    select count(*) as count_value from test group by key
  )
  group by count_value
sql

但是我正在寻找一种方法来使用Sequel命令（如果可能的话）。

我尝试了什么：

group_and_count不可链接：

#SQLite3::SQLException: no such column: count (Sequel::DatabaseError)
sel = DB[:test].group_and_count(:key).group_and_count(:count)

也无法重命名：count

sel = DB[:test].group_and_count(:key).select(:count.as(:count_value))

如果我定义了一个视图，我会得到一个DatabaseError：

DB.create_view(:count_testkey,  DB[:test].group_and_count(:key))
sel = DB[:count_testkey].group_and_count(:count)

#SQLite3::SQLException: aggregate functions are not allowed in the GROUP BY clause (Sequel::DatabaseError)
puts  Sequel::PrettyTable.string(sel)

Answer 1

我找到了解决方案：

sel = DB[:test].select(:count_value, Sequel::SQL::Function.new(:count, Sequel::Dataset::WILDCARD).as(:count)).group_by(:count_value)
.from( 
    DB[:test].select(Sequel::SQL::Function.new(:count, Sequel::Dataset::WILDCARD).as(:count_value)
        ).group_by(:key)
  )

puts sel.sql.downcase
puts  Sequel::PrettyTable.string(sel)

这会导致（SQL命令'由我精心打印）：

select `count_value`, count(*) as 'count' from (
  select count(*) as 'count_value' from `test` group by `key`) as 't1' 
group by `count_value`
+-----+-----------+
|count|count_value|
+-----+-----------+
|    1|         19|
|    1|         20|
|    1|         21|
|    1|         22|
|    1|         23|
|    1|         24|
+-----+-----------+

基于此解决方案，我定义了一种新方法Sequel::Dataset#group_and_countcount。

这是一个完整的例子：

require 'sequel'
Sequel.extension :pretty_table
DB = Sequel.sqlite()

DB.create_table( :test ){
  add_field :key,    :type => :nvarchar, :size => 10
  add_field :value, :type => :nvarchar, :size => 10
}

'a'.upto('f'){|a|
    a.upto('x'){|b|
      DB[:test].insert( :key => a, :value => b)
      DB[:test].insert( :key => 'x', :value => b) if a == 'a' #force one double count
    }
}

puts "Test DB[:test].group_and_count(:key)"
sel = DB[:test].group_and_count(:key)
puts  Sequel::PrettyTable.string(sel) 

module Sequel
  class Dataset
    def group_and_countcount(key)
      self.unfiltered.select(:count_value, Sequel::SQL::Function.new(:count, Sequel::Dataset::WILDCARD).as(:count)).group_by(:count_value)
        .from( 
            self.select(Sequel::SQL::Function.new(:count, Sequel::Dataset::WILDCARD).as(:count_value)
                ).group_by(key)
          )
    end #group_and_countcount(key)
  end
end

puts "Test DB[:test].group_and_countcount(:key)"
sel = DB[:test].group_and_countcount(:key)
puts  Sequel::PrettyTable.string(sel)

结果：

Test DB[:test].group_and_count(:key)
+-----+---+
|count|key|
+-----+---+
|   24|a  |
|   23|b  |
|   22|c  |
|   21|d  |
|   20|e  |
|   19|f  |
|   24|x  |
+-----+---+
Test DB[:test].group_and_countcount(:key)
+-----+-----------+
|count|count_value|
+-----+-----------+
|    1|         19|
|    1|         20|
|    1|         21|
|    1|         22|
|    1|         23|
|    2|         24|
+-----+-----------+

如何计算Sequel :: Dataset＃group_and_count的计数值

实施例

我尝试了什么：

1 个答案: