Conditional expressions let you use if … elif …\nelse logic within filters, annotations, aggregations, and updates. A\nconditional expression evaluates a series of conditions for each row of a\ntable and returns the matching result expression. Conditional expressions can\nalso be combined and nested like other expressions.
Nós estaremos usando o seguinte modelo nos exemplos subsequentes:
\nfrom django.db import models\n\nclass Client(models.Model):\n REGULAR = 'R'\n GOLD = 'G'\n PLATINUM = 'P'\n ACCOUNT_TYPE_CHOICES = [\n (REGULAR, 'Regular'),\n (GOLD, 'Gold'),\n (PLATINUM, 'Platinum'),\n ]\n name = models.CharField(max_length=50)\n registered_on = models.DateField()\n account_type = models.CharField(\n max_length=1,\n choices=ACCOUNT_TYPE_CHOICES,\n default=REGULAR,\n )\nWhenLink para este cabeçalho#A When() object is used to encapsulate a condition and its result for use\nin the conditional expression. Using a When() object is similar to using\nthe filter() method. The condition can\nbe specified using field lookups,\nQ objects, or Expression\nobjects that have an output_field that is a\nBooleanField. The result is provided using the\nthen keyword.
Alguns exemplos:
\n>>> from django.db.models import F, Q, When\n>>> # String arguments refer to fields; the following two examples are equivalent:\n>>> When(account_type=Client.GOLD, then='name')\n>>> When(account_type=Client.GOLD, then=F('name'))\n>>> # You can use field lookups in the condition\n>>> from datetime import date\n>>> When(registered_on__gt=date(2014, 1, 1),\n... registered_on__lt=date(2015, 1, 1),\n... then='account_type')\n>>> # Complex conditions can be created using Q objects\n>>> When(Q(name__startswith="John") | Q(name__startswith="Paul"),\n... then='name')\n>>> # Condition can be created using boolean expressions.\n>>> from django.db.models import Exists, OuterRef\n>>> non_unique_account_type = Client.objects.filter(\n... account_type=OuterRef('account_type'),\n... ).exclude(pk=OuterRef('pk')).values('pk')\n>>> When(Exists(non_unique_account_type), then=Value('non unique'))\nTenha em mente que cada um desses valores pode ser uma expressão.
\n\nCaseLink para este cabeçalho#A Case() expression is like the if … elif …\nelse statement in Python. Each condition in the provided\nWhen() objects is evaluated in order, until one evaluates to a\ntruthful value. The result expression from the matching When() object\nis returned.
Um exemplo:
\n>>>\n>>> from datetime import date, timedelta\n>>> from django.db.models import Case, CharField, Value, When\n>>> Client.objects.create(\n... name='Jane Doe',\n... account_type=Client.REGULAR,\n... registered_on=date.today() - timedelta(days=36))\n>>> Client.objects.create(\n... name='James Smith',\n... account_type=Client.GOLD,\n... registered_on=date.today() - timedelta(days=5))\n>>> Client.objects.create(\n... name='Jack Black',\n... account_type=Client.PLATINUM,\n... registered_on=date.today() - timedelta(days=10 * 365))\n>>> # Get the discount for each Client based on the account type\n>>> Client.objects.annotate(\n... discount=Case(\n... When(account_type=Client.GOLD, then=Value('5%')),\n... When(account_type=Client.PLATINUM, then=Value('10%')),\n... default=Value('0%'),\n... output_field=CharField(),\n... ),\n... ).values_list('name', 'discount')\n<QuerySet [('Jane Doe', '0%'), ('James Smith', '5%'), ('Jack Black', '10%')]>\nCase() accepts any number of When() objects as individual arguments.\nOther options are provided using keyword arguments. If none of the conditions\nevaluate to TRUE, then the expression given with the default keyword\nargument is returned. If a default argument isn’t provided, None is\nused.
If we wanted to change our previous query to get the discount based on how long\nthe Client has been with us, we could do so using lookups:
>>> a_month_ago = date.today() - timedelta(days=30)\n>>> a_year_ago = date.today() - timedelta(days=365)\n>>> # Get the discount for each Client based on the registration date\n>>> Client.objects.annotate(\n... discount=Case(\n... When(registered_on__lte=a_year_ago, then=Value('10%')),\n... When(registered_on__lte=a_month_ago, then=Value('5%')),\n... default=Value('0%'),\n... output_field=CharField(),\n... )\n... ).values_list('name', 'discount')\n<QuerySet [('Jane Doe', '5%'), ('James Smith', '0%'), ('Jack Black', '10%')]>\nCase() also works in a filter() clause. For example, to find gold\nclients that registered more than a month ago and platinum clients that\nregistered more than a year ago:
>>> a_month_ago = date.today() - timedelta(days=30)\n>>> a_year_ago = date.today() - timedelta(days=365)\n>>> Client.objects.filter(\n... registered_on__lte=Case(\n... When(account_type=Client.GOLD, then=a_month_ago),\n... When(account_type=Client.PLATINUM, then=a_year_ago),\n... ),\n... ).values_list('name', 'account_type')\n<QuerySet [('Jack Black', 'P')]>\nConditional expressions can be used in annotations, aggregations, filters,\nlookups, and updates. They can also be combined and nested with other\nexpressions. This allows you to make powerful conditional queries.
\nDigamos que queremos mudar o account_type dos nossos clientes que correspondam às datas de registro. Podemos fazer isso usando uma expressão condicional e o método update():
>>> a_month_ago = date.today() - timedelta(days=30)\n>>> a_year_ago = date.today() - timedelta(days=365)\n>>> # Update the account_type for each Client from the registration date\n>>> Client.objects.update(\n... account_type=Case(\n... When(registered_on__lte=a_year_ago,\n... then=Value(Client.PLATINUM)),\n... When(registered_on__lte=a_month_ago,\n... then=Value(Client.GOLD)),\n... default=Value(Client.REGULAR)\n... ),\n... )\n>>> Client.objects.values_list('name', 'account_type')\n<QuerySet [('Jane Doe', 'G'), ('James Smith', 'R'), ('Jack Black', 'P')]>\nWhat if we want to find out how many clients there are for each\naccount_type? We can use the filter argument of aggregate\nfunctions to achieve this:
>>> # Create some more Clients first so we can have something to count\n>>> Client.objects.create(\n... name='Jean Grey',\n... account_type=Client.REGULAR,\n... registered_on=date.today())\n>>> Client.objects.create(\n... name='James Bond',\n... account_type=Client.PLATINUM,\n... registered_on=date.today())\n>>> Client.objects.create(\n... name='Jane Porter',\n... account_type=Client.PLATINUM,\n... registered_on=date.today())\n>>> # Get counts for each value of account_type\n>>> from django.db.models import Count\n>>> Client.objects.aggregate(\n... regular=Count('pk', filter=Q(account_type=Client.REGULAR)),\n... gold=Count('pk', filter=Q(account_type=Client.GOLD)),\n... platinum=Count('pk', filter=Q(account_type=Client.PLATINUM)),\n... )\n{'regular': 2, 'gold': 1, 'platinum': 3}\nThis aggregate produces a query with the SQL 2003 FILTER WHERE syntax\non databases that support it:
SELECT count('id') FILTER (WHERE account_type=1) as regular,\n count('id') FILTER (WHERE account_type=2) as gold,\n count('id') FILTER (WHERE account_type=3) as platinum\nFROM clients;\nOn other databases, this is emulated using a CASE statement:
SELECT count(CASE WHEN account_type=1 THEN id ELSE null) as regular,\n count(CASE WHEN account_type=2 THEN id ELSE null) as gold,\n count(CASE WHEN account_type=3 THEN id ELSE null) as platinum\nFROM clients;\nThe two SQL statements are functionally equivalent but the more explicit\nFILTER may perform better.
When a conditional expression returns a boolean value, it is possible to use it\ndirectly in filters. This means that it will not be added to the SELECT\ncolumns, but you can still use it to filter results:
>>> non_unique_account_type = Client.objects.filter(\n... account_type=OuterRef('account_type'),\n... ).exclude(pk=OuterRef('pk')).values('pk')\n>>> Client.objects.filter(~Exists(non_unique_account_type))\nIn SQL terms, that evaluates to:
\nSELECT ...\nFROM client c0\nWHERE NOT EXISTS (\n SELECT c1.id\n FROM client c1\n WHERE c1.account_type = c0.account_type AND NOT c1.id = c0.id\n)\n