I have an Event model
that has start_at
time and end_at
time in my schedule app and I want to validate the overlapping time before saving.
I create my rails app on Cloud9.
My view image as followings;
Day1
07:00 - 07:20 event1
10:30 - 11:30 event2
15:40 - 16:10 event3
[add event button]
Day2
08:15 - 09:05 event4
12:08 - 13:04 event5
14:00 - 14:25 event6
[add event button]
[save schedule button]
start_at
time and end_at
time can be changed and added at the same time.
What I'd like to do is to display error if I try to add (or change to) 07:05 - 07:30
for Day1
, for example, 13:50 - 14:30
for Day2
and so on.
For example;
app_development=# select * from events;
id | start_at | end_at | title | detail | schedule_id | created_at | updated_at
----+----------+----------+--------+--------+-----------------+----------------------------+----------------------------
1 | 07:00:00 | 07:20:00 | event1 | | 1 | 2016-04-12 05:28:44.166827 | 2016-04-12 12:52:07.682872
2 | 10:30:00 | 11:30:00 | event2 | | 1 | 2016-04-12 05:28:44.17747 | 2016-04-12 12:52:07.689934
3 | 15:40:00 | 16:10:00 | event3 | | 1 | 2016-04-12 05:29:07.5005 | 2016-04-12 12:52:07.693477
I added 07:05 - 07:30
above table, but the validation doesn't work.
Although I asked the similar question, I was advised to use postgresql instead of sqlite3.
So I managed to configure postgresql, but the result is the same. It would be appreciated if you could give me how to check and display error.
schema.rb
create_table "events", force: :cascade do |t|
t.time "start_at"
t.time "end_at"
t.string "title"
t.integer "room_id"
...
create_table "rooms", force: :cascade do |t|
t.string "room"
t.integer "schedule_id"
...
create_table "schedules", force: :cascade do |t|
t.string "title"
t.integer "user_id"
t.date "departure_date"
...
Give the following models:
class Event < ActiveRecord::Base
belongs_to :room, inverse_of: :events
has_one :schedule, autosave: false, through: :room
...
validate :cannot_overlap_another_event
def cannot_overlap_another_event
range = Range.new start_at, end_at
overlaps = Event.exclude_self(id).in_range(range)
overlap_error unless overlaps.empty?
end
scope :in_range, -> range {
where('(start_at BETWEEN ? AND ?)', range.first, range.last)
}
scope :exclude_self, -> id { where.not(id: id) }
def overlap_error
errors.add(:overlap_error, 'There is already an event scheduled in this hour!')
end
class Schedule < ActiveRecord::Base
belongs_to :user
has_many :rooms, inverse_of: :schedule
accepts_nested_attributes_for :rooms, allow_destroy: true
...
class Room < ActiveRecord::Base
belongs_to :schedule, inverse_of: :rooms
has_many :events, inverse_of: :room
accepts_nested_attributes_for :events, allow_destroy: true
...
_schedule_form.html.erb
<%= render 'shared/error_messages', object: f.object %>
<%= f.label :title %>
<%= f.text_field :title, class: 'form-control' %>
<br>
<%= f.label :departure_date %>
<div class="input-group date" id="datetimepicker">
<%= f.text_field :departure_date, :value => (f.object.departure_date if f.object.departure_date), class: 'form-control' %>
<span class="input-group-addon">
<span class="glyphicon glyphicon-calendar"></span>
</span>
</div>
<script type="text/javascript">
$(function () {
$('#datetimepicker').datetimepicker({format:'YYYY-MM-DD'});
});
</script>
<br>
<div id="room">
<%= f.simple_fields_for :rooms do |a| %>
<div id="room_<%= a.object.object_id %>">
<p class="day-number-element-selector"><b>Day <%= a.index.to_i + 1 %></b></p>
<%= a.simple_fields_for :events do |e| %>
<span class="form-inline">
<p>
<%= e.input :start_at, label: false %> -
<%= e.input :end_at, label: false %>
</p>
</span>
<%= e.input :title, label: false %>
<% end %>
</div>
<%= a.link_to_add "Add event", :events, data: {target: "#room_#{a.object.object_id}"}, class: "btn btn-primary" %>
<%= a.input :room %>
<% end %>
</div>
It would be appreciated if you could give me how to check and display error.
EDIT
edit as followings;
event.rb
scope :in_range, -> range {
where('(start_at BETWEEN ? AND ? OR end_at BETWEEN ? AND ?) OR (start_at <= ? AND end_at >= ?)', range.first, range.last, range.first, range.last, range.first, range.last)
}
Although it seems to work, this validate doesn't work when I add event
on a different day as following id=8
. (see created_at
and updated_at
)
app_development=# select * from events;
id | start_at | end_at | title | detail | room_id | created_at | updated_at
----+----------+----------+--------+--------+-----------------+----------------------------+----------------------------
1 | 07:00:00 | 07:20:00 | event1 | | 1 | 2016-04-12 05:28:44.166827 | 2016-04-12 12:52:07.682872
2 | 10:30:00 | 11:30:00 | event2 | | 1 | 2016-04-12 05:28:44.17747 | 2016-04-12 12:52:07.689934
3 | 15:40:00 | 16:10:00 | event3 | | 1 | 2016-04-12 05:29:07.5005 | 2016-04-12 12:52:07.693477
8 | 07:05:00 | 07:10:00 | event4 | | 1 | 2016-04-15 21:37:58.569868 | 2016-04-15 21:39:27.956737
Your scope is headed in the right direct, but doesn't cover all your cases.
In your example, you end up checking
start_at BETWEEN 7:05 AND 7:30
, butstart_at
on Day 1 is7:00
, which is outside that range.There are four cases you need to handle:
Looking, you can see that the first three cases are handled by checking if
Then you just need to catch the fourth case by adding
You could add a similar check on
end_at
for code symmetry, but it's not strictly necessary.You don't need to reinvent the wheel in PostgreSQL, there are two implemented simple methods to achieve overlap checks:
OVERLAPS
operator:Simple enough,
However this enables one range to be exactly after the other
(in other words, it checks start <= time < end).
&&
(overlaps) operator:This is simple too, usually. But PostgreSQL doesn't have a built-in range type for
time
(however there aretsrange
,tstzrange
, anddaterange
for the other temporal types).You need to create this range type for yourself:
But after this, you can check overlapping with
Pros of range types:
you can control yourself, how do you want to handle range boundaries
f.ex. you could use
timerange(start_at, end_at, '[]')
to include both the start & the end-point of the ranges. By default it includes the start, but excludes the end-point of the ranges.it can be indexed, f.ex. with
Exclusion constraints: this is essentially the same, what you want to achieve, but it will be enforced at DB level (like,
UNIQUE
or any other constraints):