Multiple Inserts across multiple tables

David wrote:
> On 4 June, 12:50, Captain Paralytic wrote:
>> On Jun 4, 10:37 am, David wrote:
>>
>>> On 4 June, 09:27, Captain Paralytic wrote:
>>>> On Jun 3, 7:28 pm, Lennart Jonsson
>>>> wrote:> On 2010-06-03 17:25, Jerry Stuckle wrote:
>>>>> [...]
>>>>>> It's quite common to have multiple clients with one company. Different
>>>>>> departments, for instance, may have separate accounts. But billing
>>>>>> still goes to the same place.
>>>>>> And each account may have one or more users authorized to use the
>>>>>> account; each user would have to be inserted.
>>>>> All agreed, but I don't think I would end up with a data model like the
>>>>> OP's, no matter what
>>>> An author who likes playing with language (e.g. Terry Pratchett or
>>>> Douglas Adams) would say something like:
>>>> "There is no such thing as a perfect data model and this is an
>>>> excellent example."
>>> I can see everyone who has comments suggests that this is an imperfect
>>> model. I would welcome any feedback on why its imperfect, and what
>>> everyone else would do instead.
>> I did that in my first post in this thread.
>> If you need a 1:m relationship, you do it using the primary key from
>> the main (1) table and using it as a foreign key in any (m) tables.
>> If you have a 1:1 relationship, then you either put all the data in a
>> single table or, if you have a good reason for spreading the data
>> amongst many tables, you use the same primary key value in all of
>> them.
>>
>> Really all you have to do is to follow the standard normalisation
>> rules:http://delicious.com/Captain_Paralytic/normalization
>
> Ok, Now I am really confused, because i thought that is what i am
> doing?
>
> insert an address, get the last_insert_id and then update the account
> table with that id.
> I do the same for the contacts, email addresses and numbers
>
> So if im being a bit thick forgive me, but isnt that what you are
> talking about?
>
>
>

You have a 1:m relationship (client:address). How do you indicate
multiple addresses if you are inserting the address id in the client table?

Rather, client_id should be a member of address, and nothing need be
updated in client.

--
==================
Remove the "x" from my email address
Jerry Stuckle
JDS Computer Training Corp.
jstucklex.DeleteThis@attglobal.net
==================

David wrote:
> On 4 June, 13:13, Jerry Stuckle wrote:
>> David wrote:
>>> On 4 June, 12:50, Captain Paralytic wrote:
>>>> On Jun 4, 10:37 am, David wrote:
>>>>> On 4 June, 09:27, Captain Paralytic wrote:
>>>>>> On Jun 3, 7:28 pm, Lennart Jonsson
>>>>>> wrote:> On 2010-06-03 17:25, Jerry Stuckle wrote:
>>>>>>> [...]
>>>>>>>> It's quite common to have multiple clients with one company. Different
>>>>>>>> departments, for instance, may have separate accounts. But billing
>>>>>>>> still goes to the same place.
>>>>>>>> And each account may have one or more users authorized to use the
>>>>>>>> account; each user would have to be inserted.
>>>>>>> All agreed, but I don't think I would end up with a data model like the
>>>>>>> OP's, no matter what
>>>>>> An author who likes playing with language (e.g. Terry Pratchett or
>>>>>> Douglas Adams) would say something like:
>>>>>> "There is no such thing as a perfect data model and this is an
>>>>>> excellent example."
>>>>> I can see everyone who has comments suggests that this is an imperfect
>>>>> model. I would welcome any feedback on why its imperfect, and what
>>>>> everyone else would do instead.
>>>> I did that in my first post in this thread.
>>>> If you need a 1:m relationship, you do it using the primary key from
>>>> the main (1) table and using it as a foreign key in any (m) tables.
>>>> If you have a 1:1 relationship, then you either put all the data in a
>>>> single table or, if you have a good reason for spreading the data
>>>> amongst many tables, you use the same primary key value in all of
>>>> them.
>>>> Really all you have to do is to follow the standard normalisation
>>>> rules:http://delicious.com/Captain_Paralytic/normalization
>>> Ok, Now I am really confused, because i thought that is what i am
>>> doing?
>>> insert an address, get the last_insert_id and then update the account
>>> table with that id.
>>> I do the same for the contacts, email addresses and numbers
>>> So if im being a bit thick forgive me, but isnt that what you are
>>> talking about?
>> You have a 1:m relationship (client:address). How do you indicate
>> multiple addresses if you are inserting the address id in the client table?
>>
>> Rather, client_id should be a member of address, and nothing need be
>> updated in client.
>>
>
> Oh, so im doing it the wrong way round.
>

Yes. You way, how would you specify multiple addresses for a single
company?

--
==================
Remove the "x" from my email address
Jerry Stuckle
JDS Computer Training Corp.
jstucklex RemoveThis @attglobal.net
==================