> Keywords: DataAccWG
>
>
> Maria A. Nieto-Santisteban wrote:
>>
>>>>>-- id bigint unsigned auto_increment
>>>>>-- * a bigint unsigned will definitely give you the range but the
>>>>> thing
>>>>>
>>>>>--    is you don't need such an ID.
>>>>>-- * your actual key is a composed key [fovid, amp]
>>>
>>>dropped composite key from merged schema
>>
>>
>> Dropping the composite key doesn't not solve the issue.
>>
>> What I mean is you don't need an extra ID as primary key. Your primary
>> key
>> is the composed key [fovid, amp]
>
> Good you brought that up, that was my impression too,
> but we were in a hurry and there was no time to follow up.
> Can you clarify directly with Sergei whether he has any
> issues with removing the "extra" ID?

You lost me guys - let's recap:
we have key on id;
we have composite key on (amp,fovid).
I thought the comment was: the composite key is a better primary key, and id
key is "extra".

My reply was, if composite key is dropped, then the issue goes away -- it
cannot be primary key anymore. Did I misread the comment?

S.
_______________________________________________
LSST-data mailing list
[email protected]
http://www.lsstmail.org/mailman/listinfo/lsst-data

Reply via email to