Hi,
What’s a good approach for referencing lookup/reference data (static)
that
one has loaded in a reference table. Say for example “tax codes”, which
pretty much have just the database ID and then a description field.
Some
options that come to mind:
1 - By ID - however this makes an assumption that the data always loads
into
a database and gets the same IDs (i.e. when you’re loading your static
reference data)?
2 - Search by description to identify the record - but obviously if one
changes the description this isn’t very good.
3 - Perhaps create a specific identify/key as a new column that you use,
but
is not the database Rails ID? How’s this sound?
Hi,
What’s a good approach for referencing lookup/reference data (static) that
one has loaded in a reference table. Say for example “tax codes”, which
pretty much have just the database ID and then a description field. Some
options that come to mind:
1 - By ID - however this makes an assumption that the data always loads into
a database and gets the same IDs (i.e. when you’re loading your static
reference data)?
If the descriptions change, you should do it just like this.
2 - Search by description to identify the record - but obviously if one
changes the description this isn’t very good.
That’s the simpler way to solve the issue, if the descriptions do not
change. But this isn’t really “beautiful” from the database point of
view, as your data isn’t going to be normalized and I usually prefer
to work with normalized data unless denormalization is needed.
3 - Perhaps create a specific identify/key as a new column that you use, but
is not the database Rails ID? How’s this sound?
I think this is a way to complex way to solve a simple problem
if one uses the ID’s then can I assume you should then “force” the ID’s
to
be set as you want them in the migration? i.e. can’t rely on a
standard
rails migration where the database uses it auto-increment?
On Mon, Dec 22, 2008 at 7:25 AM, MaurÃcio Linhares <
On Dec 21, 4:25 pm, “Maurício Linhares” [email protected]
wrote:
[…]
3 - Perhaps create a specific identify/key as a new column that you use, but
is not the database Rails ID? How’s this sound?
I think this is a way to complex way to solve a simple problem
I totally disagree. I think this is probably the best way to do it in
certain cases – namely, when you know that the key is not going to
change. And then you might as well use that key as the primary key,
instead of an autogenerated key.
A good example of this might be a table of countries. If the data is
like this:
countries:
code: US
name: United States
code: CA
name: Canada
code: MX
name: Mexico
…then it is perfectly reasonable, in order to avoid the
unpredictability of an autogenerated primary key, to find countries by
code (method 3), and it will work far more reliably than method 2 and
more readably than method 1. It’s not overcomplicated at all. It may
even be worthwhile to use code as the primary key column, forgoing an
autogenerated key. (In this case, I think I’d still use the
autogenerated key – country abbreviations can change – but you
need to look at your data and decide what suits it best.)
