我想查询一系列表格中的数据。但是表名是在创建表的日期之后命名的,例如:2017-01-10
,2017-01-11
等。经过一些研究后我发现TABLE_DATE_RANGE
功能可能适合我的需要但是,表名必须采用<prefix><date>
格式,例如:
SELECT *
FROM TABLE_DATE_RANGE([mydataset.sea_weather_],
TIMESTAMP("2016-05-01"),
TIMESTAMP("2016-05-09"))
问题再次是我想要查询的表没有带前缀,但我仍然希望从一系列表/日期中进行查询。请指教。
答案 0 :(得分:0)
经过一些研究后我发现TABLE_DATE_RANGE功能可能符合我的需要,但表格名称必须采用
的格式<prefix><date>
我认为你的发现不正确
请尝试使用以下内容查询上周表格(假设表格名称为YYYMMDD格式 - 因此没有前缀):
for BigQuery Legacy SQL
#legacySQL
SELECT <list of fields ...>
FROM (TABLE_DATE_RANGE([yourProject:yourDataset.],
DATE_ADD(CURRENT_TIMESTAMP(), -7 - DAYOFWEEK(CURRENT_TIMESTAMP()) + 2, 'DAY'),
DATE_ADD(CURRENT_TIMESTAMP(), -1 - DAYOFWEEK(CURRENT_TIMESTAMP()) + 2, 'DAY')
))
GROUP BY 1, 2, 3
ORDER BY 1, 2, 3
您可以使用[githubarchive:day]公共数据集
进行测试#legacySQL
SELECT
STRFTIME_UTC_USEC(created_at, "%Y%m%d") AS d,
DAYOFWEEK(created_at) AS dow,
DAYOFWEEK(CURRENT_TIMESTAMP()) AS cdow,
COUNT(1) AS rows
FROM (TABLE_DATE_RANGE([githubarchive:day.],
DATE_ADD(CURRENT_TIMESTAMP(), -7 - DAYOFWEEK(CURRENT_TIMESTAMP()) + 2, 'DAY'),
DATE_ADD(CURRENT_TIMESTAMP(), -1 - DAYOFWEEK(CURRENT_TIMESTAMP()) + 2, 'DAY')
))
GROUP BY 1, 2, 3
ORDER BY 1, 2, 3
对于BigQuery Standard SQL
#standardSQL
SELECT
DATE(created_at) AS d,
EXTRACT(DAYOFWEEK FROM DATE(created_at)) AS dow,
EXTRACT(DAYOFWEEK FROM CURRENT_DATE()) AS cdow,
COUNT(1) AS c
FROM `yourProject.yourDataset.*`
WHERE _TABLE_SUFFIX
BETWEEN FORMAT_DATE('%Y%m%d',DATE_ADD(CURRENT_DATE(), INTERVAL -7 - EXTRACT(DAYOFWEEK FROM CURRENT_DATE()) + 2 DAY))
AND FORMAT_DATE('%Y%m%d',DATE_ADD(CURRENT_DATE(), INTERVAL -1 - EXTRACT(DAYOFWEEK FROM CURRENT_DATE()) + 2 DAY))
GROUP BY 1, 2, 3
ORDER BY 1, 2, 3
请注意,在标准SQL中,如果您的数据集中包含视图,则此问题可能会出现问题 - 您将收到类似Error: Views cannot be queried through prefix
的错误。
在这种情况下 - 您可以使用以下&#34;技巧&#34;例如(假设视图与那个(2 *)通配符不匹配
#standardSQL
SELECT
DATE(created_at) AS d,
EXTRACT(DAYOFWEEK FROM DATE(created_at)) AS dow,
EXTRACT(DAYOFWEEK FROM CURRENT_DATE()) AS cdow,
COUNT(1) AS c
FROM `githubarchive.day.2*`
WHERE CONCAT('2', _TABLE_SUFFIX )
BETWEEN FORMAT_DATE('%Y%m%d',DATE_ADD(CURRENT_DATE(), INTERVAL -7 - EXTRACT(DAYOFWEEK FROM CURRENT_DATE()) + 2 DAY))
AND FORMAT_DATE('%Y%m%d',DATE_ADD(CURRENT_DATE(), INTERVAL -1 - EXTRACT(DAYOFWEEK FROM CURRENT_DATE()) + 2 DAY))
GROUP BY 1, 2, 3
ORDER BY 1, 2, 3