我刚从结果集创建了分区表，以确保我理解正确。 从黑客新闻中只有三列，即time_ts，title和score。 从2013年至2015年，创建了1015个分区。 从2007年到2015年，3158个分区。（超出限制）
Query-1
Creating partitioned table based on date column.(2013-2015)
create table mydataset.y2013_y2015(title_date date, title string, score int64) partition by title_date as
select extract(date from time_ts)extracted_date,title,score from `bigquery-public-data.hacker_news.stories` where extract(date from time_ts) between '2013-01-01' and '2015-12-31'
Query-2
Creating partitioned table based on timestamp column.(2013-2015)
create table mydataset.y2013_y2015_ts(title_ts timestamp, title string, score int64) partition by date(title_ts) as
select time_ts,title,score from `bigquery-public-data.hacker_news.stories` where extract(date from time_ts) between '2013-01-01' and '2015-12-31'
Query-3
Creating partitioned table based on timestamp column.(2007-2015)
create table mydataset.y2007_y2015(ts timestamp, title string, score int64) partition by date(ts) as
select time_ts,title,score from `bigquery-public-data.hacker_news.stories` where extract(date from time_ts) between '2007-01-01' and '2015-12-31'
Error: Too many partitions produced by query, allowed 2000, query produces at least 3158 partitions
Query-1和Query-2从结果集创建分区表。
检查 - 创建了多少个分区。
select COUNT(*)No_of_partitions from [mydataset.y2013_y2015$__PARTITIONS_SUMMARY__]
select COUNT(*)No_of_partitions from [mydataset.y2013_y2015_ts$__PARTITIONS_SUMMARY__]
我说错了吗？
答案 0 :(得分：1)
是的。你是对的。 您也可以使用此代码（在标准SQL中）
#standardSQL
select count(distinct _PARTITIONDATE) from `mydataset.y2013_y2015`