如何修改R中的顶级XML节点？

Question

我想添加一个属性到XML文件的最顶层节点，然后保存该文件。我已经尝试过所有可以考虑的xpath和子集的组合，但似乎无法使其工作。用一个简单的例子：

xml_string = c(
'<?xml version="1.0" encoding="UTF-8"?>', 
'<retrieval-response status = "found">', 
     '<coredata>', 
      '<id type = "author" >12345</id>', 
     '</coredata>', 
     '<author>', 
      '<first>John</first>', 
      '<last>Doe</last>', 
     '</author>', 
'</retrieval-response>') 

# parse xml content 
xml = xmlParse(xml_string) 

当我尝试

xmlAttrs(xml["/retrieval-response"][[1]]) <- c(id = 12345) 

我得到一个错误：

object of type 'externalptr' is not subsettable 

然而，属性插入，所以我不知道我做错了。 （更多背景：这是来自Scopus API的数据的简化版本，我将数以千计的xml文件结构相似，但id在“coredata”节点，它是“作者”节点的同胞其中包含所有的数据，所以当我使用SAS将组合XML文档编译为数据集时，id和数据之间没有链接，我希望将id添加到层次结构的顶部会导致它传播到所有其他级别）。

Answer 1

编辑： 试图编辑顶部节点的方法后（见Old Answer），我意识到编辑顶层节点并不能解决我的问题，因为SAS XML映射器没有保留所有的ID。

我试着将作者id添加到每个完美工作的子节点的新方法。我还了解到，您可以使用XPath通过将它们放入一个载体，像这样选择多个节点：

c("//coredata", 
    "//affiliation-current", 
    "affiliation-history", 
    "subject-areas", 
    "//author-profile") 

所以我用最后的方案是：

files <- list.files() 

for (i in 1:length(files)) { 
    author_record <- xmlParse(files[i]) 

    xpathApply(
      author_record, c(
       "//coredata", 
       "//affiliation-current", 
       "affiliation-history", 
       "subject-areas", 
       "//author-profile" 
     ), 
      addAttributes, 
      auth_id = gsub("AUTHOR_ID:", "", xmlValue(author_record[["//dc:identifier"]])) 
    ) 

    saveXML(author_record, file = files[i]) 
} 

老答案： 经过多次实验，我发现了一个相对简单的解决方案来解决我的问题

属性可以通过简单地使用

addAttributes(xmlRoot(xmlfile), attribute = "attributeValue") 

对于我的具体情况下被添加到顶级节点，最简单的解决方案将是一个简单的循环：

setwd("C:/directory/with/individual/xmlfiles") 

files <- list.files() 

for (i in 1:length(files)) { 

author_record <- xmlParse(files[i]) 

addAttributes(node = xmlRoot(author_record), 
       id = gsub (pattern = "AUTHOR_ID:", 
           replacement = "", 
           x = xmlValue(auth[["//dc:identifier"]]) 
       ) 
) 

    saveXML(author_record, file = files[i]) 
} 

我敢肯定有更好的方法。显然我需要学习XLST，这是一个非常强大的方法！

Answer 2

为了根据数据集和数据框的结构将XML数据迁移到行和列的二维中，必须删除所有嵌套以仅迭代父级和一个子级。因此，XSLT是专门针对任何细微差别需要重构XML文档的专用声明性编程语言，它们可以用来重构XML数据以满足最终用户的需求。

给出您的示例XML，下面是一个可运行的XSLT，并将结果XML成功导入SAS。让SAS代码循环重构所有数千个XML文件。

XSLT（保存为的.xsl或.xslt格式）

<xsl:transform xmlns:xsl="http://www.w3.org/1999/XSL/Transform" version="1.0" 
     xmlns:ait="http://www.elsevier.com/xml/ani/ait" 
     xmlns:ce="http://www.elsevier.com/xml/ani/common" 
     xmlns:cto="http://www.elsevier.com/xml/cto/dtd" 
     xmlns:dc="http://purl.org/dc/elements/1.1/" 
     xmlns:ns1="http://webservices.elsevier.com/schemas/search/fast/types/v4" 
     xmlns:prism="http://prismstandard.org/namespaces/basic/2.0/" 
     xmlns:xocs="http://www.elsevier.com/xml/xocs/dtd" 
     xmlns:xoe="http://www.elsevier.com/xml/xoe/dtd" 
     exclude-result-prefixes="ait ce cto dc ns1 prism xocs xoe"> 
<xsl:output version="1.0" encoding="UTF-8" indent="yes" /> 

<xsl:template match="author-retrieval-response"> 
    <xsl:variable select="substring-after(coredata/dc:identifier, ':')" name="authorid"/> 
    <root> 
     <coredata> 
     <authorid><xsl:value-of select="$authorid"/></authorid> 
     <xsl:for-each select="coredata/*">   
      <xsl:element name="{local-name()}">  
      <xsl:value-of select="concat(.,@href)"/> 
      </xsl:element> 
     </xsl:for-each> 
     </coredata> 

     <subjectAreas> 
     <authorid><xsl:value-of select="$authorid"/></authorid> 
     <xsl:for-each select="subject-areas/*">   
      <xsl:element name="{local-name()}">  
      <xsl:value-of select="."/> 
      </xsl:element> 
     </xsl:for-each> 
     </subjectAreas> 

     <authorname> 
     <authorid><xsl:value-of select="$authorid"/></authorid> 
     <xsl:for-each select="author-profile/preferred-name/*">   
      <xsl:element name="{local-name()}">  
      <xsl:value-of select="."/> 
      </xsl:element> 
     </xsl:for-each> 
     </authorname> 

     <classifications> 
     <authorid><xsl:value-of select="$authorid"/></authorid> 
     <xsl:for-each select="author-profile/classificationgroup/classifications/*">   
      <xsl:element name="{local-name()}">  
      <xsl:value-of select="."/> 
      </xsl:element> 
     </xsl:for-each> 
     </classifications> 

     <journals> 
     <authorid><xsl:value-of select="$authorid"/></authorid> 
     <xsl:for-each select="author-profile/journal-history/journal/*">   
      <xsl:element name="{local-name()}">  
      <xsl:value-of select="."/> 
      </xsl:element> 
     </xsl:for-each> 
     </journals> 

     <ipdoc> 
     <authorid><xsl:value-of select="$authorid"/></authorid> 
     <xsl:for-each select="author-profile/affiliation-current/affiliation/ip-doc/*[not(local-name()='address')]">   
      <xsl:element name="{local-name()}">  
      <xsl:value-of select="."/> 
      </xsl:element> 
     </xsl:for-each> 
     </ipdoc> 

     <address> 
     <authorid><xsl:value-of select="$authorid"/></authorid> 
     <xsl:for-each select="author-profile/affiliation-current/affiliation/ip-doc/address/*">   
      <xsl:element name="{local-name()}">  
      <xsl:value-of select="."/> 
      </xsl:element> 
     </xsl:for-each> 
     </address> 
    </root> 
</xsl:template> 

</xsl:transform> 

SAS（使用上述脚本）

proc xsl 
    in="C:\Path\To\Original.xml" 
    out="C:\Path\To\Output.xml" 
    xsl="C:\Path\To\XSLT.xsl"; 
run; 

** STORING XML CONTENT; 
libname temp xml 'C:\Path\To\Output.xml'; 

** APPEND CONTENT TO SAS DATASETS; 
data Work.Coredata; 
    retain authorid; 
    set temp.Coredata; ** NAME OF PARENT NODE IN XML; 
run; 

data Work.SubjectAreas; 
    retain authorid; 
    set temp.SubjectAreas; ** NAME OF PARENT NODE IN XML; 
run; 

data Work.Authorname; 
    retain authorid; 
    set temp.Authorname; ** NAME OF PARENT NODE IN XML; 
run; 

data Work.Classifications; 
    retain authorid; 
    set temp.Classifications; ** NAME OF PARENT NODE IN XML; 
run; 

data Work.Journals; 
    retain authorid; 
    set temp.Journals; ** NAME OF PARENT NODE IN XML; 
run; 

data Work.Ipdoc;  
    retain authorid; 
    set temp.Ipdoc; ** NAME OF PARENT NODE IN XML; 
run; 

XML OUTPUT（其被导入作为一行和40个变量的Authorsdata数据集）

<?xml version="1.0" encoding="UTF-8"?> 
<root> 
    <coredata> 
     <authorid>1234567</authorid> 
     <url>http://api.elsevier.com/content/author/author_id/1234567</url> 
     <identifier>AUTHOR_ID:1234567</identifier> 
     <eid>9-s2.0-1234567</eid> 
     <document-count>3</document-count> 
     <cited-by-count>95</cited-by-count> 
     <citation-count>97</citation-count> 
     <link>http://api.elsevier.com/content/search/scopus?query=refauid%1234567%29</link> 
     <link>http://www.scopus.com/authid/detail.url?partnerID=HzOxMe3b&amp;authorId=1234567&amp;origin=inward</link> 
     <link>http://api.elsevier.com/content/author/author_id/1234567</link> 
     <link>http://api.elsevier.com/content/search/scopus?query=au-id%281234567%29</link> 
    </coredata> 
    <subjectAreas> 
     <authorid>1234567</authorid> 
     <subject-area>Human-Computer Interaction</subject-area> 
     <subject-area>Control and Systems Engineering</subject-area> 
     <subject-area>Software</subject-area> 
     <subject-area>Computer Vision and Pattern Recognition</subject-area> 
     <subject-area>Artificial Intelligence</subject-area> 
    </subjectAreas> 
    <authorname> 
     <authorid>1234567</authorid> 
     <initials>A.</initials> 
     <indexed-name>John A.</indexed-name> 
     <surname>John</surname> 
     <given-name>Doe</given-name> 
    </authorname> 
    <classifications> 
     <authorid>1234567</authorid> 
     <classification>1709</classification> 
     <classification>2207</classification> 
     <classification>1712</classification> 
     <classification>1707</classification> 
     <classification>1702</classification> 
    </classifications> 
    <journals> 
     <authorid>1234567</authorid> 
     <sourcetitle>Very Prestigious Journal</sourcetitle> 
     <sourcetitle-abbrev>V PRES JOU Autom</sourcetitle-abbrev> 
     <issn>10504729</issn> 
     <sourcetitle>2005 Another Prestigious Journal</sourcetitle> 
     <sourcetitle-abbrev>An. Prest. Jou. </sourcetitle-abbrev> 
    </journals> 
    <ipdoc> 
     <authorid>1234567</authorid> 
     <afnameid>Prestigious University#1111111</afnameid> 
     <afdispname>Prestigious University University</afdispname> 
     <preferred-name>Prestigious University University</preferred-name> 
     <sort-name>Prestigious University</sort-name> 
     <org-domain>pu.edu</org-domain> 
     <org-URL>http://www.pu.edu/index.shtml</org-URL> 
    </ipdoc> 
    <address> 
     <authorid>1234567</authorid> 
     <address-part>1234 Prestigious Lane</address-part> 
     <city>City</city> 
     <state>ST</state> 
     <postal-code>12345</postal-code> 
     <country>United States</country> 
    </address> 
</root> 

[R另类

由于没有全面的[R XSLT库中存在，解析将不得不在R输入语言直接完成。但是，R可以通过命令行，RCOMClient包和其他接口调用其他可执行文件（即Python，Saxon，VBA）的XSLT处理器。

尽管如此，R可以为authorid通过xmlToDataFrame()和xpathSApply()（后者类似XPath）提取XML数据：

library(XML) 

coredata <- xmlToDataFrame(nodes = getNodeSet(doc, '//coredata')) 
coredata$authorid <- gsub(pattern = "AUTHOR_ID:", replacement = "", 
          xpathSApply(doc, '//coredata/dc:identifier', xmlValue)[[1]]) 

subjectareas <- xmlToDataFrame(nodes = getNodeSet(doc, "//subject-areas")) 
subjectareas$authorid <- gsub(pattern = "AUTHOR_ID:", replacement = "", 
           xpathSApply(doc, '//coredata/dc:identifier', xmlValue)[[1]]) 

authorname <- xmlToDataFrame(nodes = getNodeSet(doc, '//author-profile/preferred-name')) 
authorname$authorid <- gsub(pattern = "AUTHOR_ID:", replacement = "", 
          xpathSApply(doc, '//coredata/dc:identifier', xmlValue)[[1]]) 

classifications <- xmlToDataFrame(nodes = getNodeSet(doc, '//author-profile/classificationgroup/classifications')) 
classifications$authorid <- gsub(pattern = "AUTHOR_ID:", replacement = "", 
           xpathSApply(doc, '//coredata/dc:identifier', xmlValue)[[1]]) 

journal <- xmlToDataFrame(nodes = getNodeSet(doc, '//author-profile/journal-history/journal')) 
journal$authorid <- gsub(pattern = "AUTHOR_ID:", replacement = "", 
         xpathSApply(doc, '//coredata/dc:identifier', xmlValue)[[1]]) 

ipdoc <- xmlToDataFrame(nodes = getNodeSet(doc, '//author-profile/affiliation-current/affiliation/ip-doc')) 
ipdoc$authorid <- gsub(pattern = "AUTHOR_ID:", replacement = "", 
         xpathSApply(doc, '//coredata/dc:identifier', xmlValue)[[1]]) 

address <- xmlToDataFrame(nodes = getNodeSet(doc, '//author-profile/affiliation-current/affiliation/ip-doc/address')) 
address$authorid <- gsub(pattern = "AUTHOR_ID:", replacement = "", 
         xpathSApply(doc, '//coredata/dc:identifier', xmlValue)[[1]])