In R XML Package, what is the difference between xmlParse and xmlTreeParse?

Question

When would I want to use the xmlParse function versus the xmlTreeParse function? Also, when are parameter values useInternalNodes=TRUE or asText=TRUE useful?

For example:

library("XML")
nct_url <- "http://clinicaltrials.gov/ct2/show/NCT00112281?resultsxml=true"
xml_doc <- xmlParse(nct_url, useInternalNodes=TRUE)

vs.

doc <- xmlTreeParse(getURL(nct_url), useInternalNodes=TRUE)
top <- xmlRoot(doc)
top[["keyword"]]
xmlValue(top[["start_date"]])
xmlValue(top[["location"]])

People seem to use the xmlTreeParse function for getting a non-repeating node via the $doc$children$... traversal. But I am not sure I understand when each approach is best. Parsing XML is one of the reasons to almost abandon R and learn Python. Lack of for-dummies examples without being forced to buy a book.

Answer 1

Here some feedback after using XML package.

• xmlParse is a version of xmlTreeParse where argument useInternalNodes is set to TRUE.
• If you want to get an R object use xmlTreeParse. This can be not very efficient and unnecessary if you want just to extract partial part of the xml document.
• If you don't want to get an R object, just a c pointer, use xmlParse. But you should know some xpath bases to manipulate the result.
• Use asText=TRUE if you have a text not a file or an url as input.

Here an example where I show the difference between the 2 functions:

txt <- "<doc>
          <el> aa </el>
       </doc>"
library(XML)
res <- xmlParse(txt,asText=TRUE)
res.tree <- xmlTreeParse(txt,asText=TRUE)

Now inspecting the 2 objects:

class(res)
[1] "XMLInternalDocument" "XMLAbstractDocument"
> class(res.tree)
[1] "XMLDocument"         "XMLAbstractDocument"

You see that res is an internal document. It is pointer to a C object. res.tree is an R object. You can get its attributes like this :

 res.tree$doc$children
$doc
<doc>
 <el>aa</el>
</doc>

For res, you should use a valid xpath request and one of theses functions ( xpathApply, xpathSApply ,getNodeSet) to inspect it. for example:

xpathApply(res,'//el')

Once you create a valid Xml Node , you can apply xmlValue, xmlGetAttr,..to extract node information. So here this 2 statements are equivalent:

## we have already an R object, just apply xmlValue to the right child
xmlValue(res.tree$doc$children$doc)
## xpathSApply create an R object and pass it to
xpathSApply(res,'//el',xmlValue)