Any reason that an order canceled still keep the completed_at value?

[jtapia]

I'm seeing that after order cancellation, the value completed_at is still populated, any reason for this? or a canceled order is considered as completed?

[seandawes]

completed_at is not what really defines the order state of complete. order state of complete would. completed_at is used to determine historical sales and other reporting as created_at and completed_at are often times lengthy date diff

[jarednorman]

Sean has this correct. You need to keep a record of when the order was completed for the purposes of sales reports, accounting, etc.

I do recognize that the whole complete? versus completed? thing is confusing.

• Spree::Order#complete? returns true if the order is currently in the complete state. If the order has been cancelled it will return false.
• Spree::Order#completed? returns true if the order ever got through checkout by checking if the completed_at timestamp is set.

While the naming is confusing, they are both useful in different contexts.

[jtapia]

Thanks for the replies and the explaining

Not sure if the concept of complete order makes sense for this case, for example:

irb(main):011:0> user
=> #<Spree::User ...>
irb(main):012:0> user.orders
=> #<ActiveRecord::Associations::CollectionProxy [#<Spree::Order ...]>
irb(main):013:0> user.orders.count
   (1.3ms)  SELECT COUNT(*) FROM "spree_orders" WHERE "spree_orders"."user_id" = $1  [["user_id", 11111]]
=> 3
irb(main):014:0> user.orders.map(&:state)
=> ["cart", "canceled", "canceled"]
irb(main):015:0> user.orders.complete
  Spree::Order Load (5.1ms)  SELECT "spree_orders".* FROM "spree_orders" WHERE "spree_orders"."user_id" = $1 AND ("spree_orders"."completed_at" IS NOT NULL)  [["user_id", 11111]]
=> #<ActiveRecord::AssociationRelation [#<Spree::Order id: ...]>
irb(main):016:0> user.orders.complete.count
   (1.3ms)  SELECT COUNT(*) FROM "spree_orders" WHERE "spree_orders"."user_id" = $1 AND ("spree_orders"."completed_at" IS NOT NULL)  [["user_id", 11111]]
=> 2
irb(main):017:0> user.orders.complete.map(&:state)
  Spree::Order Load (1.4ms)  SELECT "spree_orders".* FROM "spree_orders" WHERE "spree_orders"."user_id" = $1 AND ("spree_orders"."completed_at" IS NOT NULL)  [["user_id", 11111]]
=> ["canceled", "canceled"]

it's considering the canceled orders as completed, I mean, I would expect to see only orders on complete state

[seandawes]

You would not want to drop completed_at info as this will impact all forms of reporting. Similar confusing naming conventions occur with order numbers even as a cart has an order number but you would think an order number should only exist for an actual order. This is where the state machine really comes into play and you just need to look at everything around states.

Added complexity comes into play when you have multiple stock locations and multiple line items.

[jarednorman]

It's true that our naming here is inconsistent. The Spree::Order.complete scope returns orders that would Spree::Order#completed? would return true for, not Spree::Order#complete?.

Spree::Order.by_state("complete") will give you what you need if you need to exclude cancelled orders.