BigQuery raises exception if query with joins are not flattened - google-bigquery

Question

Query is:
select
cb.subnum as subnum,
last(
if(
(if(cu.smartcard_number is not null, 1, 0)) +
(if(rr.smart_card_number is not null, 1, 0)) > 0, 1, 0)
) as econnected_i,
from
combined.table1 as cb
left outer join each dataflow_raw_eu.table2 as cu
on cu.smartcard_number = cb.smart_card_num
left outer join each dataflow_raw_eu.table3 as rr
on rr.smart_card_number = cb.smart_card_num
group by subnum
The error is:
Error: Ambiguous field name 'imported_at' in JOIN. Please use the table qualifier before field name.
I noticed that when it runs with only one join, of either table, then the query succeeds. imported_at is a timestamp field shared by all 3 tables (the only field shared by all 3), but it is not included in the query.
If I select flatten_results in the BigQuery options, then the query succeeds; but I wish to run future queries with nested records. None of the tables in the above query have a repeated or record field.

Answer 1

Looks like this can be GBQ bug
Try below for workaround
SELECT
cb.subnum AS subnum,
LAST(
IF(
(IF(cu.smartcard_number IS NOT NULL, 1, 0)) +
(IF(rr.smart_card_number IS NOT NULL, 1, 0)) > 0, 1, 0)
) AS econnected_i,
FROM
combined.table1 AS cb
LEFT OUTER JOIN EACH (SELECT smartcard_number FROM dataflow_raw_eu.table2) AS cu
ON cu.smartcard_number = cb.smart_card_num
LEFT OUTER JOIN EACH (SELECT smart_card_number FROM dataflow_raw_eu.table3) AS rr
ON rr.smart_card_number = cb.smart_card_num
GROUP BY subnum
Please note, depends on logic and nature of data in dataflow_raw_eu.table2 and dataflow_raw_eu.table3 , you might consider using GROUP BY in subselects, like below
SELECT smartcard_number FROM dataflow_raw_eu.table2 GROUP BY smartcard_number
and
SELECT smart_card_number FROM dataflow_raw_eu.table3 GROUP BY smart_card_number