Schema advice for huge csv file

Question

Guys, I need an advice: Huge csv file (500 millions rows) to load in a table and I did. But now I need to alter the columns (they came all as varchar(50)). I'm just change one column and it's taking age...what kind of schema should I adopt? So far I applied
a simple data flow but I am wondering if I should do something like:
drop table
create table (all varchar)
data flow
alter table
No sure about

Answer 1

Is this a once off/ad-hoc load or something that'll be ongoing/BAU?
If it's ongoing then Arthur's post is the standard approach.
create a staging table with varchar(50s) or whatever. Load into that, then from that staging table go into your 'normal' table that has the correct column types.
If it's a once off, what I'd do is create a new table with the correct data types. Do a bulk insert from your table with 500mil rows.
then drop the old table and rename the new table.
Converting the columns in your 500mil table one by one is going to take a very long time, it'll be faster to do one bulk insert into a table with the correct schema
Jakub @ Adelaide, Australia Blog