docs: create table as

en_US/sql-reference/statements/create.md

@@ -55,6 +55,60 @@ ENGINE=TimeSeries
 with (ttl='10d')
 ```
 
+### Create Table From Query Result
+
+`CREATE TABLE AS` creates a table based on the results of a SELECT query. The column types and order of the new table are automatically derived from the output schema of the query.
+
+```SQL
+CREATE TABLE [database.]table_name
+(
+    [table_constraint, ...]
+)
+[PARTITION BY HASH(expr) [PARTITIONS partition_num]]
+[ENGINE = TimeSeries]
+[WITH (key=value, ...)]
+AS select_statement
+```
+
+#### Notes
+
+- Unlike regular `CREATE TABLE`, `CREATE TABLE AS` **does not allow** column definitions inside the parentheses — column types and order are derived from the SELECT statement's output schema.
+- Only table-level constraints such as `TIMESTAMP KEY(...)`, `PRIMARY KEY(...)` may appear inside the parentheses.
+- `IF NOT EXISTS` is **not supported**.
+- Only `ENGINE = TimeSeries` is currently supported.
+- Any valid `SELECT` statement can follow `AS`, including complex queries with `WHERE`, `JOIN`, etc.
+- Once the table is created, query results are automatically inserted into the new table. The return value is the number of rows written.
+
+#### Examples
+
+Basic usage
+
+```SQL
+CREATE TABLE sink (
+    timestamp key(ts)
+)
+PARTITION BY HASH(sid) PARTITIONS 1
+ENGINE=TimeSeries
+AS
+SELECT ts, sid, value FROM source WHERE sid >= 2
+```
+
+With JOIN
+
+```SQL
+CREATE TABLE sink_join (
+    timestamp key(ts)
+)
+PARTITION BY HASH(sid) PARTITIONS 1
+ENGINE=TimeSeries
+AS
+SELECT source.ts, source.sid, source.value, dim.name
+FROM source
+JOIN dim
+ON source.sid = dim.sid
+WHERE source.sid >= 2
+```
+
 ### Declare indexes in CREATE TABLE (INVERTED / VECTOR)
 
 Besides creating indexes after table creation with `CREATE INDEX`, Datalayers also supports declaring indexes directly inside `CREATE TABLE` table constraints.

zh_CN/sql-reference/statements/create.md

@@ -77,6 +77,60 @@ ENGINE=TimeSeries
 with (ttl='10d')
 ```
 
+### 从查询结果创建表
+
+`CREATE TABLE AS` 基于 SELECT 查询的结果创建表，新表的列类型和列顺序由查询的输出 schema 自动推导。
+
+```SQL
+CREATE TABLE [database.]table_name
+(
+    [table_constraint, ...]
+)
+[PARTITION BY HASH(expr) [PARTITIONS partition_num]]
+[ENGINE = TimeSeries]
+[WITH (key=value, ...)]
+AS select_statement
+```
+
+#### 说明
+
+- 与普通 `CREATE TABLE` 不同，`CREATE TABLE AS` **不允许**在括号内定义列定义。列的类型和顺序从 `SELECT` 语句的输出 schema 自动推导。
+- 括号内仅可包含表级约束，例如 `TIMESTAMP KEY(...)`、`PRIMARY KEY(...)` 等。
+- **不支持** `IF NOT EXISTS` 子句。
+- 当前仅支持 `ENGINE = TimeSeries`。
+- `AS` 后可以是任意合法的 `SELECT` 语句，支持 `JOIN`、`UNION`、CTE 等复杂查询。
+- 表创建成功后，查询结果会自动写入新表，返回值为写入的行数。
+
+#### 示例
+
+基础用法
+
+```SQL
+CREATE TABLE sink (
+    timestamp key(ts)
+)
+PARTITION BY HASH(sid) PARTITIONS 1
+ENGINE=TimeSeries
+AS
+SELECT ts, sid, value FROM source WHERE sid >= 2
+```
+
+带 JOIN 的用法
+
+```SQL
+CREATE TABLE sink_join (
+    timestamp key(ts)
+)
+PARTITION BY HASH(sid) PARTITIONS 1
+ENGINE=TimeSeries
+AS
+SELECT source.ts, source.sid, source.value, dim.name
+FROM source
+JOIN dim
+ON source.sid = dim.sid
+WHERE source.sid >= 2
+```
+
 ### 创建 Source
 
 `CREATE SOURCE` 用于定义一个流式输入对象。它描述外部数据源的字段、connector 和 format，但它本身不是一个表，无法接查询和写入。