Comparing two ManyToMany fields for same entries

Just thought I'd add another example using Python script:

Person0 = Person()
Person1 = Person()
Person2 = Person()
Person0.friends.add(Person2)
Person2.friends.add(Person0)
Person2.friends.add(Person1)
Person1.friends.add(Person2)

Now, I would like to do the following, but it seems to fail:

Person0.friends.all() in Person1.friends.all(). I would like it to
say if Person0 and Person1 share another friend in common.

Person0 and Person2 are friends
Person1 and Person2 are friends
Person0 and Person1 are NOT friends, but share a friend in common.
How does one find out that even though Person0 and Person1 are not
friends, they do share Person2 as a friend.

I can use my eye on a Python shell to see that Person2 exists on both
Peson0 and Person1, but how does one make the code see it?

Sorry for having to clarify this so much, I'm just not sure that my
last post actually explained it properly.

Thanks.

Post by Kevin
Hello,
I am building a model which shares a relation with another model
using a ManyToManyField. What I need to do, is find out which models
are on both on two seperate ManyToManyField lists. Here is a simple
friends=ManyToManyField(self)
To find out this persons direct friends, Person.friends...
To find out which friends this Person shares in common with another
Person, ????
Person1
Person6
Person3
Person8
Person2
Person6
What would be the most optimized QuerySet to find out that both
Person0 and Person1 are both friends with Person6?
Is there a specific Django app perhaps that can ease develop of this
type of data relations between objects?
Thanks.

Brett Epps

2011-10-28 18:21:52 UTC

Try this:

for friendof0 in Person0.friends.all():
for friendof1 in Person1.friends.all():
if friendof0 == friendof1:
# Person 0 and Person 1 share a friend.
else:
# They have no shared friends.

Brett

Post by Kevin
Person0 = Person()
Person1 = Person()
Person2 = Person()
Person0.friends.add(Person2)
Person2.friends.add(Person0)
Person2.friends.add(Person1)
Person1.friends.add(Person2)
Person0.friends.all() in Person1.friends.all(). I would like it to
say if Person0 and Person1 share another friend in common.
Person0 and Person2 are friends
Person1 and Person2 are friends
Person0 and Person1 are NOT friends, but share a friend in common.
How does one find out that even though Person0 and Person1 are not
friends, they do share Person2 as a friend.
I can use my eye on a Python shell to see that Person2 exists on both
Peson0 and Person1, but how does one make the code see it?
Sorry for having to clarify this so much, I'm just not sure that my
last post actually explained it properly.
Thanks.

--
You received this message because you are subscribed to the Google Groups
"Django users" group.
To unsubscribe from this group, send email to
For more options, visit this group at
http://groups.google.com/group/django-users?hl=en.

Kevin

2011-10-28 18:34:52 UTC

Thanks. Is this way good for database optimization though? I guess I
could use select_related(). I was hoping there was a QuerySet I could
run.

I made this one, although I'm not sure if it works correctly:

Person.objects.filter(name='Person1').filter(friends__friends__name='Person0')
This outputs "Person1", if I interchange Person0 with Person2, it
outputs no results. If I use Person1 in both instances, it outputs
'Person1'.
Does this QuerySet return the results I want, or should I use Brett's
suggestion?

Thanks again.

# Person 0 and Person 1 share a friend.
# They have no shared friends.
Brett

Kevin

2011-10-28 18:44:02 UTC

I think the filter I used does work to detect related friends, here is
a Pythonic example of the same code, but done differently without
filters:

Person.objects.get(name='Person1').friends.get().friends.all()

Instead of returning back 'Person1', when using filters, this returns
back 'Person0'.

Person.objects.get(name='Sean').friends.get().friends.get(name='Person0')

My next question, which would be better optimized for database hits:

* Brett's example
* using the filter version
* using the Pythonic version, but use select_related(depth=2) option?

I am trying to make sure the database hits are optimized as this will
be queries multiple times per page hit for different models. Also,
which would be faster to make sure the page loads are not burden when
lots of these types of queries go through?

Thanks.

Thanks. Is this way good for database optimization though? I guess I
could use select_related(). I was hoping there was a QuerySet I could
run.
Person.objects.filter(name='Person1').filter(friends__friends__name='Person0')
This outputs "Person1", if I interchange Person0 with Person2, it
outputs no results. If I use Person1 in both instances, it outputs
'Person1'.
Does this QuerySet return the results I want, or should I use Brett's
suggestion?
Thanks again.

# Person 0 and Person 1 share a friend.
# They have no shared friends.
Brett

Brett Epps

2011-10-28 22:35:24 UTC

I'm going to guess that my version is slow since a new query would be run
for each row returned by the initial one.

Check out an app called django-debug-toolbar. It has a view that will
show you all of the SQL queries generated as a page is built. It will
also tell you how long each query took to run. This should allow you to
compare the three methods.

Brett

Post by Kevin
I think the filter I used does work to detect related friends, here is
a Pythonic example of the same code, but done differently without
Person.objects.get(name='Person1').friends.get().friends.all()
Instead of returning back 'Person1', when using filters, this returns
back 'Person0'.
Person.objects.get(name='Sean').friends.get().friends.get(name='Person0')
* Brett's example
* using the filter version
* using the Pythonic version, but use select_related(depth=2) option?
I am trying to make sure the database hits are optimized as this will
be queries multiple times per page hit for different models. Also,
which would be faster to make sure the page loads are not burden when
lots of these types of queries go through?
Thanks.

Post by Kevin
Thanks. Is this way good for database optimization though? I guess I
could use select_related(). I was hoping there was a QuerySet I could
run.
Person.objects.filter(name='Person1').filter(friends__friends__name='Pers
on0')
This outputs "Person1", if I interchange Person0 with Person2, it
outputs no results. If I use Person1 in both instances, it outputs
'Person1'.
Does this QuerySet return the results I want, or should I use Brett's
suggestion?
Thanks again.

Post by Brett Epps
# Person 0 and Person 1 share a friend.
# They have no shared friends.
Brett

Post by Kevin
Hello,
I am building a model which shares a relation with another model
using a ManyToManyField. What I need to do, is find out which

models

Post by Kevin
are on both on two seperate ManyToManyField lists. Here is a

simple

Post by Kevin
friends=ManyToManyField(self)
To find out this persons direct friends, Person.friends...
To find out which friends this Person shares in common with another
Person, ????
Person1
Person6
Person3
Person8
Person2
Person6
What would be the most optimized QuerySet to find out that both
Person0 and Person1 are both friends with Person6?
Is there a specific Django app perhaps that can ease develop of

this

Post by Kevin
type of data relations between objects?
Thanks.

--
You received this message because you are subscribed to the Google

Groups

Post by Kevin
"Django users" group.
To unsubscribe from this group, send email to
For more options, visit this group at
http://groups.google.com/group/django-users?hl=en.

Kevin

2011-10-29 11:57:35 UTC

Your version isn't as bad you think, here is the SQL it generates:

SELECT "relations_person"."id",
"relations_person"."name"
FROM "relations_person"
INNER JOIN "relations_person_friends" ON ("relations_person"."id" =
"relations_person_friends"."to_person_id")
WHERE "relations_person_friends"."from_person_id" = 1

SELECT "relations_person"."id",
"relations_person"."name"
FROM "relations_person"
INNER JOIN "relations_person_friends" ON ("relations_person"."id" =
"relations_person_friends"."to_person_id")
WHERE "relations_person_friends"."from_person_id" = 3

It only does 2 queries, but the database's test data only has 3
Persons.

Person.objects.filter(name='Person1').filter(friends__friends__name='Person0')
makes only a single query:

SELECT "relations_person"."id",
"relations_person"."name"
FROM "relations_person"
INNER JOIN "relations_person_friends" ON ("relations_person"."id" =
"relations_person_friends"."from_person_id")
INNER JOIN "relations_person" T3 ON
("relations_person_friends"."to_person_id" = T3."id")
INNER JOIN "relations_person_friends" T4 ON (T3."id" =
T4."from_person_id")
INNER JOIN "relations_person" T5 ON (T4."to_person_id" = T5."id")
WHERE ("relations_person"."name" = Person1
AND T5."name" = Person0)LIMIT 21

I'll need to play around a little and see what I can figure out.

Post by Brett Epps
I'm going to guess that my version is slow since a new query would be run
for each row returned by the initial one.
Check out an app called django-debug-toolbar. It has a view that will
show you all of the SQL queries generated as a page is built. It will
also tell you how long each query took to run. This should allow you to
compare the three methods.
Brett

Thanks. Is this way good for database optimization though? I guess I
could use select_related(). I was hoping there was a QuerySet I could
run.
Person.objects.filter(name='Person1').filter(friends__friends__name='Pers
on0')
This outputs "Person1", if I interchange Person0 with Person2, it
outputs no results. If I use Person1 in both instances, it outputs
'Person1'.
Does this QuerySet return the results I want, or should I use Brett's
suggestion?
Thanks again.

# Person 0 and Person 1 share a friend.
# They have no shared friends.
Brett

Post by Kevin
Hello,
I am building a model which shares a relation with another model
using a ManyToManyField. What I need to do, is find out which

models

Post by Kevin
are on both on two seperate ManyToManyField lists. Here is a

simple

this