雪花无法聚类 table

Question

（代表 Snowflake 用户提交...）

---

I am having table with 88 million rows and applied clustering on date [having 28 distinct values] but when i checked clustering info and clustering depth , snowflake putting all data in single partition and clustering ratio is zero. I followed snowflake documentation for creating clustered table and inserted data into table. What can be the issue ? Any step i missed which is not part of documentation ? Do we need to enable clustering at account level ?

---

近期推荐：

For customers that predated automatic clustering, you must manually enable reclustering with alter table t1 resume recluster; otherwise reclustering is enabled simply by setting the cluster key.

First, check the column AUTO_CLUSTERING_ON from SHOW TABLES;

Next, check for results under AUTOMATIC_CLUSTERING_HISTORY();

---

还有其他建议或创新的 "outside-of-the-box" 解决方案吗？？谢谢！

Answer 1

您可能需要让您的帐户 SE 为您的帐户打开自动集群。特别是如果您在集群历史视图中没有看到任何操作。

select * from  "SNOWFLAKE"."ACCOUNT_USAGE"."AUTOMATIC_CLUSTERING_HISTORY";

您可以通过在您的帐户 'billing and usage' 下查找 'automatic_clustering' 仓库来验证这一点，假设您具有 ACCOUNTADMIN 角色访问权限。如果你没有看到这个特殊的内部仓库，很可能它没有打开。

AUTOMATIC_CLUSTERING

如果您已经尝试恢复 recluster 但它仍然关闭，很可能它没有在帐户级别打开。

旁注 8800 万行应该不足以聚集，但我想这取决于您的流失或更新。您始终可以创建 table as 并自己排序数据，以查看是否可以更好地消除分区。

Create Table with new order:

create table newtable as select * from oldtable order by date;

check the partition count with:

select system$clustering_information('tablename','(date)');