Eu primeiro uniria as tabelas (no BigQuery a sintaxe para união é vírgula). Existem então duas abordagens:
- Use as funções analíticas FIRST_VALUE e LAST_VALUE.
SELECT id, timestamp_first, timestamp_last, data FROM (SELECT id, timestamp, FIRST_VALUE(timestamp) OVER( PARTITION BY id ORDER BY timestamp ASC ROWS BETWEEN UNBOUNDED PRECEDING AND UNBOUNDED FOLLOWING) AS timestamp_first, LAST_VALUE(timestamp) OVER( PARTITION BY id ORDER BY timestamp ASC ROWS BETWEEN UNBOUNDED PRECEDING AND UNBOUNDED FOLLOWING) AS timestamp_last FROM table1, table2, table3
- Use a agregação MIN/MAX no carimbo de data/hora para encontrar o primeiro/último e, em seguida, junte-se novamente às mesmas tabelas.
SELECT a.id id, timestamp_first, timestamp_last, data FROM (SELECT id, data FROM table1,table2,table3) a INNER JOIN (SELECT id, MIN(timestamp) timestamp_first, MAX(timestamp) timestamp_last FROM table1,table2,table3 GROUP BY id) b ON a.id = b.id