Re: r there downsides to explicitly naming a pk column xxxx_pk

it 4am here right now sorry :-)

i'm talking about the name of the column holding the data. I mentioned the
fk constraint naming part only because we've decided on suffixing the
constraint name with an "fk", and i guess that led to his suggestion of
suffixing a table's pk column with "pk". for example: we have a table named
"units" (for units of measurement); its pk currently is named unit_id. he
wanted to name it "unit_pk".

On Fri, Dec 15, 2017 at 4:22 AM, David G. Johnston <
david(dot)g(dot)johnston(at)gmail(dot)com> wrote:

> On Thu, Dec 14, 2017 at 1:14 PM, john snow <ofbizfanster(at)gmail(dot)com> wrote:
>
>> instead of the more conventional xxxx_id or just id?
>>
>> sorry if this may be a foolish question to some, but i'm trying to think
>> thru
>> a junior colleagues's proposal. the discussion occurred while we were
>> discussing naming our foreign key constraints using the convention
>> "childtable_parenttable_colname_fk".
>>
>
> ​Are you talking about the constraint name or the name of the column
> holding the data?​
>
> ​Identifiers in PostgreSQL can only be 64 characters (bytes?) long.
>
> If it is the column name I wouldn't get too crazy or people writing out
> SQL joins manually will be asking you to pay their medical bills...
>
> I generally avoid naming any column "id" - tables get short code aliases
> and those prefix the "id". I then name the column in the FK the exact same
> name. I rely on system defaults for choosing the names of the
> corresponding constraints and indexes.
>
> David J.
>
>