如何使用Spark XML创建嵌套标签

Question

我正在尝试将df编写为xml文件。我的问题是如何在xml输出中执行分组依据。

我的输入数据作为csv文件：

user_id|acnt_id|transaction|desc 
1      |1234|   012345| desc1
1      |1234|   102345| desc2
1      |5678|   123454| desc1
2      |1234|   345672| desc1

我正在读取csv文件并将其写为xml文件，我想知道在编写xml时如何指定嵌套模式。

目前我能得到的

<transactions>
  <transaction>
          <user_id>1</user_id>
          <acnt_id>1234</acnt_id>
          <transaction>012345</transaction>
          <desc>desc1</desc>
  </transaction>
  <transaction>
          <user_id>1</user_id>
          <acnt_id>1234</acnt_id>
          <transaction>102345</transaction>
          <desc>desc2</desc>
  </transaction>
  <transaction>
          <user_id>1</user_id>
          <acnt_id>5678</acnt_id>
          <transaction>123454</transaction>
          <desc>desc1</desc>
  </transaction>
  <transaction>
          <user_id>2</user_id>
          <acnt_id>1234</acnt_id>
          <transaction>345672</transaction>
          <desc>desc2</desc>
  </transaction>
</transactions>

预期o / p

<transactions>
  <transaction>
          <user_id>1</user_id>
              <acnt_id>1234</acnt_id>
                  <transaction>012345</transaction>
                  <desc>desc1</desc>
                  <transaction>102345</transaction>
                  <desc>desc2</desc>
              <acnt_id>5678</acnt_id>
                  <transaction>123454</transaction>        
                  <desc>desc1</desc>
  </transaction>
  <transaction>
          <user_id>2</user_id>
            <acnt_id>1234</acnt_id>
              <transaction>345672</transaction>
              <desc>desc2</desc>
  </transaction>
</transactions>