which way of querying is better?

let's say you have a database with a table filled with an addressid (yes i need an id for addresses), names and addresses... "fname", "lname", "address1", "city", "state", "zip"...

now, my question is... is it better to
1) make a column of address2 (for po box's, apt #'s and such) in the same table where many of the "address2" fields will be blank, or
2) to make a whole new table with the "addressid" and the "address2" and create a query where it would have the database go through both tables

thanx : )

Whichever creates the least amount of replicated data.

table.user
id
fname
lname
...

table.address
id
user_id
street
city
state
zip
...

one-to-many: one user has many addresses.

thanx, but actually.. i meant address2 as in:
name -> John Doe
street1 -> 123 elm st
city -> new york
state -> new york
zip -> 11001

address2 (or street2) would be apt #101

since many people live in houses that don't need a second line (apt #101) for their address, would it be better create a second table which holds all of the address2 or just keep it in the same table and save (what i'm guessing) processing power from going through another table
or is a repeated "null" count as replicated data?
thanx

oh gawsh no

put address2 in the same table and let it go null

rudy
http://r937.com/

haha... k.. thanx.... but can you give me a reason why i shouldn't create another table? too much extra work for something so small? takes up too much processing power?
thanx.. : )

queries are simpler

do not go by processing speed and disk space

that's what we did in the 80's, when those things cost something

nowadays the difference in speed when you join two tables (assuming you wanted fields out of both tables, which you wouldn't, necessarily, always, in your case), is negligible

and the space difference is a wash, too

however, when it comes to writing queries, go ahead, have a go yourself, write the queries to insert an address, update an address, and delete an address -- be fair, do two sets of queries, one for the single table model, and the other set for the two-table model

then come back and you tell us which way is easier

haha.. k... thanx.. : )