你可以遍历xml-tree并收集具有匹配名称的节点：

def ins = '''
<root>
    <stock>
        <field1>Header</field1>
        <field2>H1</field2>
        <positions>
            <data1>Hello</data1>
            <data2>P1</data2>
        </positions>
        <positions>
            <data1>World</data1>
            <data2>P2</data2>
        </positions>
    </stock>
    <stock>
        <field1>Header</field1>
        <field2>H2</field2>
        <positions>
            <data1>Hello</data1>
            <data2>P3</data2>
        </positions>
        <positions>
            <data1>World</data1>
            <data2>P4</data2>
        </positions>
    </stock>
</root>
'''
def content = new XmlSlurper().parseText(ins)

def csv = content.depthFirst()
        .findAll { it.name().matches(/(field|data).*/) }
        .collect{ it.text() }.collate(2)
        .collect { it.join("|") }.join("\n")

或者，如果您需要一种可以扩展字段和数据标记数量的通用方法：

def csv = content.depthFirst()
        .findAll { it.name() in ["stock", "positions"] }
        .collect { node -> node.children()
                    .findAll{ it.name() != "positions" }
                    .collect{ it.text() } }
        .collect { it.join("|") }.join("\n")

它不是超级高效的，因为它将所有内容都读入内存，但这将工作：

String ​xmlString = """<stocks><stock>             <field1>Header</field1>                 <field2>H1</field2>         <positions>               <data1>Hello</data1>               <data2>P1</data2>         </positions>         <positions>               <data1>World</data1>               <data2>P2</data2>         </positions>     </stock>     <stock>             <field1>Header</field1>                 <field2>H2</field2>         <positions>               <data1>Hello</data1>               <data2>P3</data2>         </positions>         <positions>               <data1>World</data1>               <data2>P4</data2>         </positions>     </stock>​</stocks>"""​​​
def xml = new ​XmlSlurper​().parseText( xmlString )

List<List<String>> output = []
xml.stock.each { stock ->
   output << [stock.field1.text(),stock.field2.text()]
   stock.positions.each { position ->
      output << [ position.data1.text(), position.data2.text() ]
   }
}
println( output.collect { it.join("|") }.join("\n") )

def csv = content.depthFirst()
    .findAll { it.name() in ["stock", "positions"] }
    .collect { node -> node.children()
                .findAll{ it.name() != "positions" }
                .collect{ it.text() } }
    .collect { it.join("|") }.join("\n")

是否也有一种方法可以动态地获取根和主体的XML注解，而不直接通过名称来指示它们？