hive lateral view语句

朴灿烈づ我的快乐病毒、 2022-04-05 07:53 189阅读 0赞

原文地址：https://cwiki.apache.org/confluence/display/Hive/LanguageManual+LateralView\#

lateral view用于和split, explode等UDTF一起使用，它能够将一列数据拆成多行数据，在此基础上可以对拆分后的数据进行聚合。

一个简单的例子，假设我们有一张表pageAds，它有两列数据，第一列是pageid string，第二列是adid\_list，即用逗号分隔的广告ID集合：

<table> 
 <tbody> 
  <tr> 
   <td>string pageid</td> 
   <td>Array&lt;int&gt; adid_list</td> 
  </tr> 
  <tr> 
   <td>"front_page"</td> 
   <td>[1, 2, 3]</td> 
  </tr> 
  <tr> 
   <td>"contact_page"</td> 
   <td>[3, 4, 5]</td> 
  </tr> 
 </tbody> 
</table>

要统计所有广告ID在所有页面中出现的次数。

首先分拆广告ID：

> 1.  `SELECT pageid, adid`
> 2.  `FROM pageAds LATERAL VIEW explode(adid_list) adTable AS adid;`

执行结果如下：

<table> 
 <tbody> 
  <tr> 
   <td>string pageid</td> 
   <td>int adid</td> 
  </tr> 
  <tr> 
   <td>"front_page"</td> 
   <td>1</td> 
  </tr> 
  <tr> 
   <td>"front_page"</td> 
   <td>2</td> 
  </tr> 
  <tr> 
   <td>"front_page"</td> 
   <td>3</td> 
  </tr> 
  <tr> 
   <td>"contact_page"</td> 
   <td>3</td> 
  </tr> 
  <tr> 
   <td>"contact_page"</td> 
   <td>4</td> 
  </tr> 
  <tr> 
   <td>"contact_page"</td> 
   <td>5</td> 
  </tr> 
 </tbody> 
</table>

接下来就是一个聚合的统计：

> 1.  `SELECT adid, count(1)`
> 2.  `FROM pageAds LATERAL VIEW explode(adid_list) adTable AS adid`
> 3.  `GROUP BY adid;`

<table> 
 <tbody> 
  <tr> 
   <td>int adid</td> 
   <td>count(1)</td> 
  </tr> 
  <tr> 
   <td>1</td> 
   <td>1</td> 
  </tr> 
  <tr> 
   <td>2</td> 
   <td>1</td> 
  </tr> 
  <tr> 
   <td>3</td> 
   <td>2</td> 
  </tr> 
  <tr> 
   <td>4</td> 
   <td>1</td> 
  </tr> 
  <tr> 
   <td>5</td> 
   <td>1</td> 
  </tr> 
 </tbody> 
</table>

多个lateral view语句： 一个FROM语句后可以跟多个lateral view语句，后面的lateral view语句能够引用它前面的所有表和列名。 以下面的表为例：

<table> 
 <tbody> 
  <tr> 
   <td>Array&lt;int&gt; col1</td> 
   <td>Array&lt;string&gt; col2</td> 
  </tr> 
  <tr> 
   <td>[1, 2]</td> 
   <td>[a", "b", "c"]</td> 
  </tr> 
  <tr> 
   <td>[3, 4]</td> 
   <td>[d", "e", "f"]</td> 
  </tr> 
 </tbody> 
</table>

> 1.  `SELECT myCol1, col2 FROM baseTable`
> 2.  `LATERAL VIEW explode(col1) myTable1 AS myCol1;`

的执行结果为：

<table> 
 <tbody> 
  <tr> 
   <td>int mycol1</td> 
   <td>Array&lt;string&gt; col2</td> 
  </tr> 
  <tr> 
   <td>1</td> 
   <td>[a", "b", "c"]</td> 
  </tr> 
  <tr> 
   <td>2</td> 
   <td>[a", "b", "c"]</td> 
  </tr> 
  <tr> 
   <td>3</td> 
   <td>[d", "e", "f"]</td> 
  </tr> 
  <tr> 
   <td>4</td> 
   <td>[d", "e", "f"]</td> 
  </tr> 
 </tbody> 
</table>

加上一个lateral view：

> 1.  `SELECT myCol1, myCol2 FROM baseTable`
> 2.  `LATERAL VIEW explode(col1) myTable1 AS myCol1`
> 3.  `LATERAL VIEW explode(col2) myTable2 AS myCol2;`

它的执行结果为：

<table> 
 <tbody> 
  <tr> 
   <td>int myCol1</td> 
   <td>string myCol2</td> 
  </tr> 
  <tr> 
   <td>1</td> 
   <td>"a"</td> 
  </tr> 
  <tr> 
   <td>1</td> 
   <td>"b"</td> 
  </tr> 
  <tr> 
   <td>1</td> 
   <td>"c"</td> 
  </tr> 
  <tr> 
   <td>2</td> 
   <td>"a"</td> 
  </tr> 
  <tr> 
   <td>2</td> 
   <td>"b"</td> 
  </tr> 
  <tr> 
   <td>2</td> 
   <td>"c"</td> 
  </tr> 
  <tr> 
   <td>3</td> 
   <td>"d"</td> 
  </tr> 
  <tr> 
   <td>3</td> 
   <td>"e"</td> 
  </tr> 
  <tr> 
   <td>3</td> 
   <td>"f"</td> 
  </tr> 
  <tr> 
   <td>4</td> 
   <td>"d"</td> 
  </tr> 
  <tr> 
   <td>4</td> 
   <td>"e"</td> 
  </tr> 
  <tr> 
   <td>4</td> 
   <td>"f"</td> 
  </tr> 
 </tbody> 
</table>

注意上面语句中，两个lateral view按照出现的次序被执行。