BigQuery Unnest based on a condition + cross join - google-bigquery

Question

Consider the following schema for monthly purchases, where items and payment are nested columns:
user STRING
items RECORD
class STRING
spent INT
payment RECORD
method STRING
e.g. MWE (nested items & payment)
User items payment
Joe class spent method
fruit 45 Direct Debit
drinks 10 Credit Card
fish 20
drugs 35
I'd like to unnest both the field items and payment but in different ways:
Restrict the field 'items' to those ones containing class='fruit' or class='drink' and unstack the values into columns
Regular cross join user - payment methods
One possible solution would be to split into two steps:
with
payments as (
select user, pay.method
from table
cross join table.payments as pay
)
, items_fruit as (
select user, items.spent as spent_fruit
from table
cross join table.items as items
where items.class IN ('fruit')
)
, items_drinks as (
select user, items.spent as spent_drinks
from table
cross join table.items as items
where items.class IN ('drinks')
)
select *
from payments
INNER JOIN items_fruit using(user)
INNER JOIN items_drinks using(user)
i.e. the outcome of the operation looking like below. However, it doesn't seem elegant nor efficient, what is the best way to achieve it?
user payment spent_fruit spent_drinks
Joe Direct Debit 45 10
Joe Credit Card 45 10

Answer 1

Consider below simple approach
select * from (
select user, class, spent, method as payment
from your_table t, t.items, t.payment
)
pivot (sum(spent) as spent for class in ('fruit', 'drinks'))
if applied to sample data in your question - output is