You can't achieve it without a "custom function", but it's not that hard to
write....
//totally untested//
def search_whatever(people=[], items=[], clothings=[]):
q = db.people.name.belongs(people)
if items:
q = q & (db.item_person.person_id == db.person.id) #the reference
q = q & (db.item.name.belongs(items))
if clothing:
q = q & (
(db.clothing.id == db.clothing_person.clothing_id) &
(db.person.id == db.clothing_person.person_id)
) #the reference
q = q & (db.clothing.name.belongs(clothings))
........
rows = db(q).select(db.person.ALL)
return rows
On Wednesday, January 29, 2014 11:03:06 PM UTC+1, Apple Mason wrote:
>
> Is there a more lenient version of what_I_want that will give me based on
> what I put in? For example,
>
> If I just want all people named Bob, it would return all people named Bob.
>
> If I just want all people named Bob or nicknamed Bobcat, then I wlll get
> all people named Bob or nicknamed Bobcat.
>
> If I just want all people nicknamed Bobcat who also have item1, item2,
> then I get all people nicknamed Bobcat with item1,item2
>
> If I just want all people named Bob and nicknamed Bobcat and have item1,
> and clothing1, clothing2, then I get all people named Bob and nicknamed
> Bobcat with item1, clothing1, clothnig2
>
> Right now if I do:
>
> what_i_want = (
> (db.person.name=='Bob') &
> (db.item.name=='item1') &
> (db.clothing.name=='clothing1')
> )
>
> I get results. But if I do:
>
> what_I_want = (
> (db.person.name=='Bob')
> )
>
> It doesn't return any rows.
>
> On Wednesday, January 29, 2014 3:14:09 AM UTC-5, Niphlod wrote:
>>
>> why the hassle of using joins like those ones ?
>> If you're not fond of searching through left joins, and you still want
>> your whole dataset "consistent", and a search "a-la-fulltext".....better do
>> something like this
>>
>> whole_set = (
>> (db.person.id == db.clothing_person.person_id) &
>> (db.clothing.id == db.clothing_person.clothing_id) &
>> (db.item_person.person_id == db.person.id) &
>> (db.item_person.item_id == db.item.id)
>> )
>>
>> then, you can search it as
>>
>> what_I_want = (
>> (db.person.name == 'Bob') &
>> ....
>> (db.item.name == 'item1')
>> )
>>
>> rows = db(whole_set)(what_I_want).select()
>>
>> On Wednesday, January 29, 2014 6:29:10 AM UTC+1, Apple Mason wrote:
>>>
>>> I want to search across some many to many tables, but with certain
>>> conditions.
>>>
>>>
>>> db.define_table('person',
>>> Field('name', 'string'),
>>> Field('nickname', 'string'))
>>>
>>> db.define_table('clothing',
>>> Field('name', 'string'))
>>>
>>> db.define_table('item',
>>> Field('name', 'string'))
>>>
>>> db.define_table('item_person',
>>> Field('person_id', 'reference person'),
>>> Field('item_id', 'reference item'))
>>>
>>> db.define_table('clothing_person',
>>> Field('person_id', 'reference person'),
>>> Field('clothing_id', 'reference clothing'))
>>>
>>>
>>>
>>> How would I find all people who have the name 'Bob' or nickname 'Bobcat'
>>> AND have items called 'item1' and 'item2' AND have clothing 'clothing1' ?
>>>
>>> For example, these are valid results:
>>>
>>> Bob has item1, item2 and clothing1
>>> Bobcat has item1, item2 and clothing1
>>>
>>> Would I use a join for this? Maybe something like:
>>>
>>> db( (db.person.name.like('Bob')) |
>>> (db.person.name.like('Bobcat')).select(db.person.ALL, join=[
>>> db.item_person.on( (db.item.id==db.item_person.item_id)
>>> & ((db.item.name=='item1') & (db.item.name='item2'))),
>>> db.clothing_person.on(
>>> (db.clothing.id==db.clothing_person.clothing_id)
>>> & (db.clothing.name=='clothing1'))
>>> ])
>>>
>>> But that doesn't seem correct.
>>>
>>
--
Resources:
- http://web2py.com
- http://web2py.com/book (Documentation)
- http://github.com/web2py/web2py (Source code)
- https://code.google.com/p/web2py/issues/list (Report Issues)
---
You received this message because you are subscribed to the Google Groups
"web2py-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to web2py+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
