BigQuery:删除分区表中的重复项

Question

我有一个按插入时间划分的BQ表。我正试图从表中删除重复项。这些是真正的重复：对于2个重复的行，所有列都是相等的-当然，拥有一个唯一的键可能会有所帮助：-(

起初，我尝试了一个SELECT查询来枚举重复项并将其删除：

SELECT
* EXCEPT(row_number)
FROM (
SELECT
*,
ROW_NUMBER() OVER (PARTITION BY id_column) row_number
FROM
`mytable`)
WHERE
row_number = 1

这会产生唯一的行，但会创建一个不包括分区数据的新表，所以这并不好。

我在这里看到了这个答案，它指出保留分区的唯一方法是使用上面的查询逐个遍历分区，并保存到特定的目标表分区。

我真正想做的是使用DMLDELETE来删除重复的行。我尝试了类似于这个答案所暗示的东西：

DELETE
FROM `mytable` AS d
WHERE (SELECT ROW_NUMBER() OVER (PARTITION BY id_column)
FROM `mytable ` AS d2
WHERE d.id = d2.id) > 1;

但接受的答案不起作用，导致BQ错误：

Error: Correlated subqueries that reference other tables are not supported unless they can be de-correlated, such as by transforming them into an efficient JOIN

如果有人能提供一种更简单的(DML或其他(方法来处理这件事，那就太好了，这样我就不需要单独遍历所有分区了。

Answer 1

这是一种破解，但您可以使用MERGE语句删除表的所有内容，并仅原子地重新插入不同的行。这里有一个例子：

-- Create a table with some duplicate rows
CREATE TABLE dataset.PartitionedTable
PARTITION BY date AS
SELECT x, CONCAT('foo', CAST(x AS STRING)) AS y, DATE_SUB(CURRENT_DATE(), INTERVAL x DAY) AS date
FROM UNNEST(GENERATE_ARRAY(1, 10)) AS x, UNNEST(GENERATE_ARRAY(1, 10));

现在对于MERGE部分：

-- Execute a MERGE statement where all original rows are deleted,
-- then replaced with new, deduplicated rows:
MERGE dataset.PartitionedTable AS t1
USING (SELECT DISTINCT * FROM dataset.PartitionedTable) AS t2
ON FALSE
WHEN NOT MATCHED BY TARGET THEN INSERT ROW
WHEN NOT MATCHED BY SOURCE THEN DELETE

Answer 2

您可以在一条SQL MERGE语句中完成此操作，而无需创建额外的表。

-- WARNING: back up the table before this operation
-- FOR large size timestamp partitioned table 
-- -------------------------------------------
-- -- To de-duplicate rows of a given range of a partition table, using surrage_key as unique id
-- -------------------------------------------
DECLARE dt_start DEFAULT TIMESTAMP("2019-09-17T00:00:00", "America/Los_Angeles") ;
DECLARE dt_end DEFAULT TIMESTAMP("2019-09-22T00:00:00", "America/Los_Angeles");
MERGE INTO `gcp_project`.`data_set`.`the_table` AS INTERNAL_DEST
USING (
SELECT k.*
FROM (
SELECT ARRAY_AGG(original_data LIMIT 1)[OFFSET(0)] k 
FROM `gcp_project`.`data_set`.`the_table` AS original_data
WHERE stamp BETWEEN dt_start AND dt_end
GROUP BY surrogate_key
)
) AS INTERNAL_SOURCE
ON FALSE
WHEN NOT MATCHED BY SOURCE
AND INTERNAL_DEST.stamp BETWEEN dt_start AND dt_end -- remove all data in partiion range
THEN DELETE
WHEN NOT MATCHED THEN INSERT ROW

信用：https://gist.github.com/hui-zheng/f7e972bcbe9cde0c6cb6318f7270b67a