BigQuery 如何根据重叠间隔连接 2 个表？

Question

I need to convert this join from Postgres to BigQuery:

-- postgres
SELECT *
FROM table_a
INNER JOIN table_b
ON tstzrange(table_a.start, table_a.end) && tstzrange(table_b.start, table_b.end)

I've tried the 4 overlapping cases but in BigQuery it doesn't work like that. My result always ends up missing some windows.

INNER JOIN table_b
ON 
-- case 1
a.start <= b.start AND b.end <= a.end
OR
-- case 2
a.start <= b.start AND b.start <= a.end
OR 
-- case 3
b.start <= a.start AND a.end <= b.end
OR
-- case 4
b.start <= a.start AND a.start <= b.end

Answer 1

Consider below condition to check if there is an overlapped period.

• a.start <= b.end AND b.start <= a.end

sample query is

WITH table_a AS (
  SELECT TIMESTAMP '2023-01-01 09:00:00' start, TIMESTAMP '2023-01-10 09:00:00' `end` UNION ALL
  SELECT '2023-01-05 09:00:00' start, '2023-01-15 09:00:00' `end` 
),
table_b AS (
  SELECT TIMESTAMP '2022-12-15 09:00:00' start, TIMESTAMP '2023-01-04 09:00:00' `end` UNION ALL 
  SELECT '2023-01-12 09:00:00' start, '2023-01-20 09:00:00' `end`
)
SELECT * FROM table_a a JOIN table_b b ON a.start <= b.end AND b.start <= a.end;

Query results

Answer 2

The issue was the NULL value is automatically handled as lower and upper bound in Postgres tstzrange . It's not handled in basic queries or in BQ. Using COALESCE() solved the issue.