在R中是否有更好的方法来完成这个XML抓取任务?

Question

我有一些XML看起来像:

<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE plist PUBLIC "-//Apple Computer//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
<plist version="1.0">
<array>
    <dict>
    <key>bbox.NE.lat</key>
    <string>-27.45433</string>
    <key>bbox.NE.lon</key>
    <string>153.01474</string>
    <key>bbox.SW.lat</key>
    <string>-27.45706</string>
    <key>bbox.SW.lon</key>
    <string>153.01239</string>
    <key>crs</key>
    <string>EPSG 4326</string>
    <key>found</key>
    <string>1</string>
</dict>
<array>
    <dict>
        <key>bbox</key>
        <dict>
            <key>bbox.NE.lat</key>
            <string>-27.45433</string>
            <key>bbox.NE.lon</key>
            <string>153.01474</string>
            <key>bbox.SW.lat</key>
            <string>-27.45706</string>
            <key>bbox.SW.lon</key>
            <string>153.01239</string>
        </dict>
        <key>centroid</key>
        <dict>
            <key>lat</key>
            <dict>
                <key>lat</key>
                <string>-27.45513</string>
                <key>lon</key>
                <string>153.0137</string>
            </dict>
        </dict>
        <key>id</key>
        <string>33037721</string>
        <key>properties</key>
        <dict>
            <key>amenity</key>
            <string>university</string>
            <key>name</key>
            <string>Queensland University of Technology</string>
            <key>osm_element</key>
            <string>way</string>
            <key>osm_id</key>
            <string>26303436</string>
        </dict>
    </dict>
</array>
</array>
</plist>

为了再现性，我用一个循环抓取XML:

# initialize the vector
queries<-(0)
# loop over coordinates - signups$latlong is just a vector of coordinates
# signups$latlong[1] = "51.5130004883%20-0.1230000034 
# the space between lat & long is URL encoded
for(i in 1:length(signups$latlong)){
#self-imposed rate-limiting
Sys.sleep(0.05)
# if query returns an error, write NA and move on, also output to console so I can keep an eye on it
if(class(try(queries[i]<- getURL(paste("http://geocoding.cloudmade.com/[API KEY HERE]/geocoding/v2/find.plist?object_type=university&around=",signups$latlong[i],"&results=5&distance=closest", sep="")), silent=T))=="try-error")
  {
    queries[i]<-NA
    print("NA")
    print(i)
    }
  # just a progress indicator
  else(print(i))
}

，我只是在大学的名字后面。我对XML了解不多，但我已经弄清楚了这是可行的:

pagetree<-(0)
for (i in 1:length(queries)){
  plist<-xmlTreeParse(queries[i])$doc$children$plist
  nodes<-getNodeSet(plist, "//array//array//dict")
  if(class(try(pagetree[i]<-xmlValue(nodes[[5]][[4]]), silent = T))=="try-error") pagetree[i]<-NA
}

然而，我看到提到了xmlXPathApply()，我想知道解析循环是否不能重写以使用它。

我的问题是我不知道如何为我想要的部分编写XPath，因为有多个具有相同名称的节点。

不要犹豫，建议一个更好的方法来做错误处理，以及

Answer 1

修复xml后…

用xpath遍历它的方法是:

> plist = xmlParse('data.xml')
> xpathSApply(plist, '/plist/array/dict/dict/string', xmlValue)
[1] "-27.45433"                           "153.01474"                          
[3] "-27.45706"                           "153.01239"                          
[5] "university"                          "Queensland University of Technology"
[7] "way"                                 "26303436" 

您可以像正常一样索引输出。

但是，如果节点有属性，例如<string type='uniname">...</string>，那么您可以使用"@"语法，如:

> xpathSApply(plist, '/plist/array/dict/dict/string[@type='uniname']', xmlValue)

另一种可能更适合此plist格式的方法是:

> sapply(getNodeSet(plist, '//key[text() = "name"]'), function(x) xmlValue(getSibling(x)))
[1] "Queensland University of Technology"