Question

我正在寻找一种在xml数据透视查询中包含非聚合值的方法。我的代码如下所示：

select * from (
select country, month from travel
where <parameters, as far as i know irrelevant to the discussion>
pivot xml (
count(month) travellers,
for country in (select country from table_a where <parameters>)
)

查询表格，其中每一行代表一个人旅行，这样月份的数量将返回该月份的旅行人数。

此查询返回以下xml：

<PivotSet>
    <item>
        <column name = "COUNTRY">AE</column>
        <column name = "TRAVELLERS">1</column>
    </item>
    <item>
        <column name = "COUNTRY">AL</column>
        <column name = "TRAVELLERS">1</column>
    </item>
    <item>
        <column name = "COUNTRY">BE</column>
        <column name = "TRAVELLERS">2</column>
    </item>
    <item>
    ...
    </item>
</PivotSet>

我想要的是获得b和c的非聚合值，我希望它们看起来像：

select * from (
select country, month from travel
where <parameters, irrelevant to the discussion>
pivot xml (
count(month) travellers,
for country, month in (select country, month from table_a where <parameters>)
)

但这不起作用，因为pivot似乎只允许一列的非聚合值。

我希望看到以下xml返回：

<PivotSet>
    <item>
        <column name="MONTH">JANUARY</column>
        <column name = "COUNTRY">AE</column>
        <column name = "TRAVELLERS">1</column>
    </item>
    <item>
        <column name="MONTH">FEBRUARY</column>
        <column name = "COUNTRY">AL</column>
        <column name = "TRAVELLERS">1</column>
    </item>
    <item>
        <column name="MONTH">MARCH</column>
        <column name = "COUNTRY">BE</column>
        <column name = "TRAVELLERS">2</column>
    </item>
    <item>
    ...
    </item>
</PivotSet>

我有什么选择让这项工作成功？我误解了枢轴功能是如何工作的吗？

Answer 1

就我在问题中的代码片段中所看到的，忽略我认为是你引入的错误所引发的错误;你只需要在你正在转动的两列上添加括号：

select *
from (
  select country, month from travel
  where <parameters, as far as i know irrelevant to the discussion>
)
pivot xml (
  count(*) travellers
  for (country, month) in (select country from table_a where <parameters>)
)

The pivot clause syntax diagram表示您可以在括号中使用单个表达式或逗号分隔的表达式列表。没有括号，它不会期望第二个表达式（, month）并抛出“ORA-01738：缺少IN关键字”，因为它预期接下来要看到的内容。

使用CTE进行演示，并假设您为了简单起见，从同一张表中获取了不同的月份/国家/地区：

-- CTE for dummy data with travel (country, month) as ( select 'US', date '2017-01-01' from dual connect by level <= 3 union all select 'US', date '2017-02-01' from dual connect by level <= 1 union all select 'France', date '2017-01-01' from dual connect by level <= 1 union all select 'France', date '2017-02-01' from dual connect by level <= 5 union all select 'Spain', date '2017-01-01' from dual connect by level <= 5 union all select 'Spain', date '2017-02-01' from dual connect by level <= 2 ) -- actual query select xmlserialize(document country_month_xml indent size=2) from ( select country, month from travel ) pivot xml ( count(*) travellers for (country, month) in (select distinct country, month from travel) );

得到：

<PivotSet> <item> <column name="COUNTRY">France</column> <column name="MONTH">2017-01-01</column> <column name="TRAVELLERS">1</column> </item> <item> <column name="COUNTRY">France</column> <column name="MONTH">2017-02-01</column> <column name="TRAVELLERS">5</column> </item> <item> <column name="COUNTRY">Spain</column> <column name="MONTH">2017-01-01</column> <column name="TRAVELLERS">5</column> </item> <item> <column name="COUNTRY">Spain</column> <column name="MONTH">2017-02-01</column> <column name="TRAVELLERS">2</column> </item> <item> <column name="COUNTRY">US</column> <column name="MONTH">2017-01-01</column> <column name="TRAVELLERS">3</column> </item> <item> <column name="COUNTRY">US</column> <column name="MONTH">2017-02-01</column> <column name="TRAVELLERS">1</column> </item> </PivotSet>

我添加了一个XMLSerialize调用只是为了让它更具可读性......

带有其他（非聚合）列的Oracle SQL xml pivot

1 个答案: