This is an automated email from the ASF dual-hosted git repository.
kassiez pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/doris-website.git
The following commit(s) were added to refs/heads/master by this push:
new 00d1873e44c update subquery document (#1422)
00d1873e44c is described below
commit 00d1873e44c72b033854d296642d3ee4fe7ee642
Author: starocean999 <li...@selectdb.com>
AuthorDate: Tue Dec 10 20:35:42 2024 +0800
update subquery document (#1422)
## Versions
- [x] dev
- [x] 3.0
- [ ] 2.1
- [ ] 2.0
## Languages
- [x] Chinese
- [x] English
## Docs Checklist
- [ ] Checked by AI
- [ ] Test Cases Built
---
docs/query-data/subquery.md | 14 +++++++-------
.../current/query-data/subquery.md | 14 +++++++-------
.../version-3.0/query-data/subquery.md | 14 +++++++-------
versioned_docs/version-3.0/query-data/subquery.md | 14 +++++++-------
4 files changed, 28 insertions(+), 28 deletions(-)
diff --git a/docs/query-data/subquery.md b/docs/query-data/subquery.md
index 9e46b6eae3a..78d7618a81d 100644
--- a/docs/query-data/subquery.md
+++ b/docs/query-data/subquery.md
@@ -304,20 +304,20 @@ With this flag, the `where` filtering condition can be
rewritten as `where mark_
## Common Issues
-Since the output of a scalar subquery must be a single value, Doris adopts
different processing methods for correlated and non-correlated scalar
subqueries.
+Since the output of a scalar subquery must be a single value, a runtime error
will be reported when the subquery returns more than one row of data.
### For Correlated Scalar Subqueries
-Currently, Doris can only statically ensure that the subquery outputs a single
value (i.e., a single aggregate function without `group by`). Therefore, when
using correlated scalar subqueries, a `group by`-less aggregate function such
as `any_value` needs to be added according to requirements, so that the
optimizer can recognize the single-value semantics smoothly. Users need to
ensure that the subquery always returns only one value. If the subquery
actually returns multiple values (othe [...]
+When using a correlated quantifier subquery, if the subquery that satisfies
the correlation condition returns more than one row of data, a runtime error
will be reported.
Please refer to the following SQL example:
```sql
--- Correlated scalar subquery lacking a single aggregate function without
group by, currently not supported
-select t1.*, (select t2.c1 from t2 where t1.c2 = t2.c2) from t1;
-
--- Add a single aggregate function for the optimizer to recognize smoothly
-select t1.*, (select any_value(t2.c1) from t2 where t1.c2 = t2.c2) from t1;
+-- If there are more than 1 row in the t2 table that satisfies t1.c2 = t2.c2
in the associated scalar subquery, a runtime error will be reported
+select t1.*, (select t2.c1 from t2 where t1.c2 = t2.c2) from t1;
+
+-- Example error message
+ERROR 1105 (HY000): errCode = 2, detailMessage =
(127.0.0.1)[INVALID_ARGUMENT][E33] correlate scalar subquery must return only 1
row
```
### For Non-Correlated Scalar Subqueries
diff --git
a/i18n/zh-CN/docusaurus-plugin-content-docs/current/query-data/subquery.md
b/i18n/zh-CN/docusaurus-plugin-content-docs/current/query-data/subquery.md
index 88e4c8b5b05..ee0ccb41a5c 100644
--- a/i18n/zh-CN/docusaurus-plugin-content-docs/current/query-data/subquery.md
+++ b/i18n/zh-CN/docusaurus-plugin-content-docs/current/query-data/subquery.md
@@ -302,25 +302,25 @@ where
## 常见问题
-由于标量子查询的输出必须是一个单值，Doris 对于关联和非关联的标量子查询采取了不同的处理方式。
+由于标量子查询的输出必须是一个单值，如果子查询返回的数据量超过一条记录，将会报告运行时错误。
### 对于关联的标量子查询
-目前 Doris 暂时只能以静态的方式确保子查询输出为单值（即没有 `group by`
的单个聚合函数）。因此，在使用关联标量子查询时，需要根据需求添加没有 `group by`
的聚合函数，如`any_value`，以便优化器能顺利识别单值语义。用户需要保证子查询一定只返回一个值，如果子查询实际返回多个值（其他数据库系统会在运行时报错），由于添加了聚合函数，它始终只返回一个值，虽然能得到结果，但可能和预期不符。
+在使用关联标量子查询时，如果满足关联条件的子查询返回的数据量超过一条记录，将会报告运行时错误。
请参考以下 SQL 示例：
```sql
--- 关联的标量子查询，缺少单个无 group by 的聚合函数，目前不支持
-select t1.*, (select t2.c1 from t2 where t1.c2 = t2.c2) from t1;
+-- 关联的标量子查询，如果t2表中满足t1.c2 = t2.c2的数据多于 1 条，则会报运行时错误
+select t1.*, (select t2.c1 from t2 where t1.c2 = t2.c2) from t1;
--- 添加单个聚合函数，让优化器顺利识别
-select t1.*, (select any_value(t2.c1) from t2 where t1.c2 = t2.c2) from t1;
+-- 报错信息样例如下
+ERROR 1105 (HY000): errCode = 2, detailMessage =
(127.0.0.1)[INVALID_ARGUMENT][E33] correlate scalar subquery must return only 1
row
```
### 对于非关联的标量子查询
-Doris 会在运行时添加一个`assert num rows`算子，如果子查询返回的数据大于一条，则会报一个运行时错误。
+Doris 会在运行时添加一个`assert num rows`算子，如果子查询返回的数据量超过一条记录，将会报告运行时错误。
请参考以下 SQL 示例：
diff --git
a/i18n/zh-CN/docusaurus-plugin-content-docs/version-3.0/query-data/subquery.md
b/i18n/zh-CN/docusaurus-plugin-content-docs/version-3.0/query-data/subquery.md
index 88e4c8b5b05..81378c0278f 100644
---
a/i18n/zh-CN/docusaurus-plugin-content-docs/version-3.0/query-data/subquery.md
+++
b/i18n/zh-CN/docusaurus-plugin-content-docs/version-3.0/query-data/subquery.md
@@ -302,25 +302,25 @@ where
## 常见问题
-由于标量子查询的输出必须是一个单值，Doris 对于关联和非关联的标量子查询采取了不同的处理方式。
+标量子查询必须输出单一值。若子查询返回的数据量超过一条记录，将会报告运行时错误。
### 对于关联的标量子查询
-目前 Doris 暂时只能以静态的方式确保子查询输出为单值（即没有 `group by`
的单个聚合函数）。因此，在使用关联标量子查询时，需要根据需求添加没有 `group by`
的聚合函数，如`any_value`，以便优化器能顺利识别单值语义。用户需要保证子查询一定只返回一个值，如果子查询实际返回多个值（其他数据库系统会在运行时报错），由于添加了聚合函数，它始终只返回一个值，虽然能得到结果，但可能和预期不符。
+在使用关联标量子查询时，如果满足关联条件的子查询返回的数据量超过一条记录，将会报告运行时错误。
请参考以下 SQL 示例：
```sql
--- 关联的标量子查询，缺少单个无 group by 的聚合函数，目前不支持
-select t1.*, (select t2.c1 from t2 where t1.c2 = t2.c2) from t1;
+-- 关联的标量子查询，如果t2表中满足t1.c2 = t2.c2的数据多于 1 条，则会报运行时错误
+select t1.*, (select t2.c1 from t2 where t1.c2 = t2.c2) from t1;
--- 添加单个聚合函数，让优化器顺利识别
-select t1.*, (select any_value(t2.c1) from t2 where t1.c2 = t2.c2) from t1;
+-- 报错信息样例如下
+ERROR 1105 (HY000): errCode = 2, detailMessage =
(127.0.0.1)[INVALID_ARGUMENT][E33] correlate scalar subquery must return only 1
row
```
### 对于非关联的标量子查询
-Doris 会在运行时添加一个`assert num rows`算子，如果子查询返回的数据大于一条，则会报一个运行时错误。
+Doris 会在运行时添加一个`assert num rows`算子，如果子查询返回的数据量超过一条记录，将会报告运行时错误。
请参考以下 SQL 示例：
diff --git a/versioned_docs/version-3.0/query-data/subquery.md
b/versioned_docs/version-3.0/query-data/subquery.md
index 9e46b6eae3a..78d7618a81d 100644
--- a/versioned_docs/version-3.0/query-data/subquery.md
+++ b/versioned_docs/version-3.0/query-data/subquery.md
@@ -304,20 +304,20 @@ With this flag, the `where` filtering condition can be
rewritten as `where mark_
## Common Issues
-Since the output of a scalar subquery must be a single value, Doris adopts
different processing methods for correlated and non-correlated scalar
subqueries.
+Since the output of a scalar subquery must be a single value, a runtime error
will be reported when the subquery returns more than one row of data.
### For Correlated Scalar Subqueries
-Currently, Doris can only statically ensure that the subquery outputs a single
value (i.e., a single aggregate function without `group by`). Therefore, when
using correlated scalar subqueries, a `group by`-less aggregate function such
as `any_value` needs to be added according to requirements, so that the
optimizer can recognize the single-value semantics smoothly. Users need to
ensure that the subquery always returns only one value. If the subquery
actually returns multiple values (othe [...]
+When using a correlated quantifier subquery, if the subquery that satisfies
the correlation condition returns more than one row of data, a runtime error
will be reported.
Please refer to the following SQL example:
```sql
--- Correlated scalar subquery lacking a single aggregate function without
group by, currently not supported
-select t1.*, (select t2.c1 from t2 where t1.c2 = t2.c2) from t1;
-
--- Add a single aggregate function for the optimizer to recognize smoothly
-select t1.*, (select any_value(t2.c1) from t2 where t1.c2 = t2.c2) from t1;
+-- If there are more than 1 row in the t2 table that satisfies t1.c2 = t2.c2
in the associated scalar subquery, a runtime error will be reported
+select t1.*, (select t2.c1 from t2 where t1.c2 = t2.c2) from t1;
+
+-- Example error message
+ERROR 1105 (HY000): errCode = 2, detailMessage =
(127.0.0.1)[INVALID_ARGUMENT][E33] correlate scalar subquery must return only 1
row
```
### For Non-Correlated Scalar Subqueries
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscr...@doris.apache.org
For additional commands, e-mail: commits-h...@doris.apache.org
