You may want something like this:

import operator
field_lookups = (‘title__icontains’, ‘author__icontains’,)
q_list = [Q(**{f:q}) for f in field_lookups]
Model.objects.filter(reduce(operator.or_, q_list))

[1] http://docs.python.org/library/operator.html#operator.or_

One remark:

args = ( Q( title__icontains = ‘Foo’ ) | Q( title__icontains = ‘Bar’ ) )
in the example should read
args = ( Q( title__icontains = ‘Foo’ ) | Q( title__icontains = ‘Bar’ ) , )
or *args fails to unpack. Notice trailing comma.

Regards,
–
Vladimir

You should be able to just create your Q objects and then OR them together in a list before you pass them to .filter() as follows:


args = []
q1 = Q( field = blah, field2 = blahblah )
q2 = Q( field3 = blah, field4 = blahblah )
args.append( q1 | q2 )
MyModel.objects.filter( *args )


‘Q(field=blah,field2=blahblah)|Q(field3=blah,field4=blahblah)’

But I can’t seem to find a way to plug that in to the filter(). I’ve tried eval(), exec(), and compile(), but no luck. Ideas?

Your comment was very helpful. I was looking for a way to dynamically create Q objects and you provided the answer!

Also, great name!


kwargs_author = {'author__icontains': 'Vinicius'}
kwargs_title = {'title__icontains': 'Vinicius'}
Entry.objects.filter(Q(**kwargs_author) | Q(**kwargs_title))


In my case, I have a form field for each kind of search (integer, date, string, boolean, primary key, choice field). I have a search form also. So, I populate my form with the GET data and it uses a field’s method called get_criteria wich returns a dictionary with in this format (‘lookup_method’: ‘value’). I pass the search string from the GET data to the get_criteria of each field and then get the dictionary returned, and pass to the Q object.

You’re right. Q objects are really handy especially for cases where OR-ing is required. Interestingly, it turns out that QuerySet’s filter method accepts both args and kwargs. So it is possible to mix and match both Q objects and standard filter parameters together. E.g.,


kwargs = { 'deleted_datetime__isnull': True }
args = ( Q( title__icontains = 'Foo' ) | Q( title__icontains = 'Bar' ) )
entries = Entry.objects.filter( *args, **kwargs )


I am going to update this post to reflect this fact.

I like the kwargs approach a little more since it involves less keystrokes, is slightly more readable (a matter of personal taste) and the end result is essentially the same.

Must reads for anyone wondering how QuerySets work:
http://docs.djangoproject.com/en/dev/topics/db/queries/#id3
http://docs.djangoproject.com/en/dev/ref/models/querysets/#when-querysets-are-evaluated


# start with everything, or a subset with Entry.objects.filter(foo) if you like
entries = Entry.objects.all()


if exclude_deleted:
entries = entries.filter(deleted_datetime__isnull=True)


if category is not None:
entries = entries.filter(category=category)


if current_user_only:
entries = entries.filter(user=request.user)


if title_search_query != '':
entries = entries.filter(title__icontains=title_search_query)