5. `select` — choose and compute columns

From here on you have three sample DataFrames on every page — orders, customers, and events — the little e-commerce dataset this course runs on. Peek at orders whenever you like:

orders.show(5)
orders.printSchema()

select is projection: it returns a new DataFrame with exactly the columns you name — no more, no less.

orders.select("order_id", "product", "category").show(5)

Computed columns need an alias

You can put an expression in select, not just a bare name. Multiply two columns, and you've got a new one — but give it a name with alias, or you'll be stuck referring to it as (quantity * unit_price):

orders.select(
    "order_id",
    (F.col("quantity") * F.col("unit_price")).alias("line_total"),
).show(5)

F.col("quantity") is how you name a column inside an expression — that's the next lesson. For now: every computed column gets an alias. Notice a couple of line_totals come out empty — two orders have no quantity (a missing value times anything is null). We deal with nulls properly in Module 3.

Keep everything, add one

select("*", ...) keeps all existing columns and tacks yours on the end:

orders.select("*", (F.col("quantity") * F.col("unit_price")).alias("line_total")).show(3)

Your turn

Return three columns from orders — order_id, product, and the line revenue (quantity * unit_price) aliased as revenue — as result.

result = orders  # <- select order_id, product, and quantity*unit_price AS revenue
result.show()

result = orders.select(
    "order_id",
    "product",
    (F.col("quantity") * F.col("unit_price")).alias("revenue"),
)