活动记录每日计数数组wrt created_at或Updated_at

Question

I am working on a school project where I have tables class_standard and enrollments

enrollments belongs_to class_standard and class_standard has many enrollments .

enrollments are created_at in different days. I want to get daily counts of enrollments in last 30 days in an array.

ie [# of enrollments 1st day, 2nd day, 3rd day, .....]

Answer 1

This is the sample of the code that gets the daily count of Enrollments created and stores them in an array.

arr = [] #initialize an array
(0..29).each do |x| #run a loop for 30 days
  date = Time.now - x.days #get a specific day
  arr << Enrollment.where(created_at: date.midnight..date.end_of_day).count #get count of enrollments on a specific day and add this to your array
end
arr = arr.reverse #reverse the array, because you want the today's count as last item of the array

Answer 2

What you want to do is select a count grouped by day. How exactly to achieve this depends on what database is used.

Postgres:

date_trunc('day', created_at)

MySQL:

date_format(created_at, '%Y-%m-%d')

For other db's check the documentation for date functions.

Altogether the query you want is something like this:

SELECT 
  count(*) AS count, 
  date_trunc('day', created_at) AS day 
FROM "enrollments" 
WHERE (created_at >= ?) -- one month ago 
GROUP BY day 
ORDER BY day

We can write this in ActiveRecord as:

Enrollment.select("count(*) AS count, to_char(created_at, 'YYYY-MM-DD') AS day")
  .group("day")
  .order("day")
  .where(created_at: (1.month.ago..Time.current))

But this gives an ActiveRecord::Relation of records which is not very useful. To get the raw query values we want to use #select_all .

sql = Enrollment.select("count(*) AS count, date_trunc('day', created_at) AS day")
  .group("day")
  .order("day")
  .where("created_at >= ?", 1.month.ago)
  .to_sql

results = Enrollment.connection.select_all(sql)

When you convert the results to an array you will get an array of hashes:

[{"count"=>10, "day"=>"2018-10-20 00:00:00"}, {"count"=>10, "day"=>"2018-10-21 00:00:00"}, {"count"=>10, "day"=>"2018-10-22 00:00:00"}, {"count"=>10, "day"=>"2018-10-23 00:00:00"}, {"count"=>10, "day"=>"2018-10-24 00:00:00"}, {"count"=>10, "day"=>"2018-10-25 00:00:00"}, ...]

You can extract the count by mapping the array:

results.map {|r| r["count"] }

If you want the results just for one (or more) ClassStandards just add an additional condition to the where clause:

sql = Enrollment.select("count(*) AS count, date_trunc('day', created_at) AS day")
  .group("day")
  .order("day")
  .where("created_at >= ?", 1.month.ago)
  .where(class_standard: @class_standard)
  .to_sql

Or you call it off the association:

sql = @class_standard.enrollments
  .select("count(*) AS count, date_trunc('day', created_at) AS day")
  .group("day")
  .order("day")
  .where("created_at >= ?", 1.month.ago)
  .to_sql

Answer 3

If you want to get count of all days including the days where there isn't any enrollment:

daily_enrollments = []
i = 29
while i >= 0
  daily_enrollments >> Enrollments.where( created_at: Time.zone
                                                          .now
                                                          .beginning_of_day-i.day..
                                                      Time.zone
                                                          .now
                                                          .end_of_day-i.day).count

  i = i-1
end

Output Example

[0,0,0,5,4,0,0,0,5,0,23,0,0,23,0,0,1,0,11,0,21,32,44,0,0,9,6,3,22,1]

This will give you an array of all 30 days. But this isn't seems an efficient code. This will run 30 queries for each record.

What seems better to me is group the records on created_at and get records for the days which have some records. ie excluding days with no record.

Enrollments.where(created_at: Time.current - 30.days..Time.current)
           .group("date(created_at)")
           .count
           .sort
           .to_h

Output Example

{Mon, 19 Nov 2018=>12}