在第一个节点内的任何级别循环 select 遍历节点和 select 个子节点

Question

我正在使用 xml2 包。我想：

Select 个具有特定标签的节点（在本例中为所有 envelope）
循环遍历这些节点和内部的 select 个节点，而不管嵌套（在这种情况下，在 card-entry 个标签内找到的所有 value 个标签）
用 space

有三个信封，所以我希望能够 return 一个包含三个向量的向量，其中包含 card-entry 标签内的 value 标签内的所有文本。我无法做到这一点。我怎样才能做到这一点？为什么我的以下方法不起作用？

MWE

library(xml2)

myxml <- read_xml('
<inside>
 <box>
   <card-entry>
     <card-id type="integer">605383</card-id>
     <value>get well</value>
   </card-entry>
 </box>
 <envelope>
     <card-entry>
       <card-id type="integer">605380</card-id>
       <value>coke</value>
     </card-entry>
     <card-entry>
       <card-id type="integer">610954</card-id>
       <value>pizza</value>
     </card-entry>  
     <card-entry>
       <card-id type="integer">605381</card-id>
       <value>surprise</value>
     </card-entry>
     <card-entry>
       <card-id type="integer">610958</card-id>
       <value>joke</value>
       <random>true</random>
     </card-entry> 
 </envelope>
 <envelope>
     <card-entry>
       <card-id type="integer">605381</card-id>
       <value>charlie horse</value>
     </card-entry>
     <card-entry>
       <card-id type="integer">605380</card-id>
       <value>rug bug</value>
     </card-entry>
     <subenvelope>
       <value>dont get me</value>
     </subenvelope>
     <card-entry>
       <card-id type="integer">610954</card-id>
       <value>mario cart</value>
     </card-entry>  
 </envelope>
 <envelope>
     <card-entry>
       <card-id type="integer">605377</card-id>
       <value>trogdor</value>
     </card-entry>
     <subenvelope>
       <card-entry>
         <card-id type="integer"></card-id>
         <value>jorb</value>
       </card-entry>
     </subenvelope>
     <card-entry>
       <card-id type="integer">605333</card-id>
       <value></value>
     </card-entry>    
 </envelope>
</inside>
'
)

期望的输出

## [1] coke pizza surprise joke
## [2] charlie horse rug bug mario cart
## [3] trogdor jorb

我试过的

children <- lapply(xml2::xml_find_all(myxml, '//envelope'), xml2::xml_children)

rm_na <- function(x) x[!is.na(x)]

lapply(children, function(x){

    paste(rm_na(unlist(xml2::xml_text(xml2::xml_child(x, '//card-entry//value')))), collapse = ' ')

})

## [[1]]
## [1] ""
## 
## [[2]]
## [1] ""
## 
## [[3]]
## [1] ""
## 
## There were 11 warnings (use warnings() to see them)


lapply(children, function(x){

    paste(rm_na(unlist(xml2::xml_text(xml2::xml_find_all(x, '//card-entry//value')))), collapse = ' ')

})

## [[1]]
## [1] "get well coke pizza surprise joke charlie horse rug bug mario cart trogdor jorb "
## 
## [[2]]
## [1] "get well coke pizza surprise joke charlie horse rug bug mario cart trogdor jorb "
## 
## [[3]]
## [1] "get well coke pizza surprise joke charlie horse rug bug mario cart trogdor jorb "

Answer 1

在你的内部选择器上，你想让它相对于当前节点，所以你想要.//而不是//，后者会再次搜索整个树

sapply(xml_find_all(myxml,"//envelope"), function(x) 
  paste(xml_text(xml_find_all(x, ".//card-entry/value")), collapse=" ")
)

# [1] "coke pizza surprise joke"        
# [2] "charlie horse rug bug mario cart"
# [3] "trogdor jorb "

在第一个节点内的任何级别循环 select 遍历节点和 select 个子节点

Loop selected over nodes and select subnodes at any level within first node

r

xml2

MWE

期望的输出

我试过的