我有一个如下所示的XML数据集：
<protocol ID='.'>
<HEAD></HEAD>
<block ID='...'>
<HEAD></HEAD>
<trial ID='.....'>
<HEAD></HEAD>
<seq ID=''>
<HEAD></HEAD>
<calibration CLASS='affine-calibration' ID='New Calibration'>
<AX>.........</AX>
<BX>-........</BX>
<AY>.........</AY>
<BY>.........</BY>
<type>'por'</type>
</calibration>
<POR TIME='......'>
<PUPIL>.</PUPIL>
<BLINK>.</BLINK>
<V>...</V>
<H>...</H>
<PLANEINTRWV>...</PLANEINTRWV>
<PLANEINTRWH>...</PLANEINTRWH>
<PLANE>.</PLANE>
</POR>
<POR TIME='......'>
<PUPIL>.</PUPIL>
<BLINK>.</BLINK>
<V>...</V>
<H>...</H>
<PLANEINTRWV>...</PLANEINTRWV>
<PLANEINTRWH>...</PLANEINTRWH>
<PLANE>.</PLANE>
</POR>
<POR TIME='......'>
<PUPIL>.</PUPIL>
<BLINK>.</BLINK>
<V>...</V>
<H>...</H>
<PLANEINTRWV>...</PLANEINTRWV>
<PLANEINTRWH>...</PLANEINTRWH>
<PLANE>.</PLANE>
</POR>
</seq>
</trial>
<trial ID='.....'>
<HEAD></HEAD>
<seq ID=''>
<HEAD></HEAD>
<calibration CLASS='affine-calibration' ID='New Calibration'>
<AX>.........</AX>
<BX>-........</BX>
<AY>.........</AY>
<BY>.........</BY>
<type>'por'</type>
</calibration>
<POR TIME='......'>
<PUPIL>.</PUPIL>
<BLINK>.</BLINK>
<V>...</V>
<H>...</H>
<PLANEINTRWV>...</PLANEINTRWV>
<PLANEINTRWH>...</PLANEINTRWH>
<PLANE>.</PLANE>
</POR>
<POR TIME='......'>
<PUPIL>.</PUPIL>
<BLINK>.</BLINK>
<V>...</V>
<H>...</H>
<PLANEINTRWV>...</PLANEINTRWV>
<PLANEINTRWH>...</PLANEINTRWH>
<PLANE>.</PLANE>
</POR>
</seq>
</trial>
</block>
</protocol>
使用XML包，提取POR标记的子项和标记属性的最简洁方法是什么？
我把这个有效的kludge聚集在一起，但它很慢（由于最有可能的xpathSApply调用）并且难以读取。
trackToDataFrame = function(file) {
doc2=xmlParse(file)
timeStamps = t(xpathSApply(doc2, '//*[@TIME]', function(x) c(name=xmlName(x), xmlAttrs(x))))
dd2 = xmlToDataFrame(getNodeSet(doc2, "//POR"), colClasses=c(rep("integer", 7)))
dd2 = cbind(dd2, timeStamps)
dd2
}
调用数据集返回：
PUPIL BLINK V H PLANEINTRWV PLANEINTRWH PLANE name TIME
1 NA NA NA NA NA NA NA POR ......
2 NA NA NA NA NA NA NA POR ......
3 NA NA NA NA NA NA NA POR ......
4 NA NA NA NA NA NA NA POR ......
5 NA NA NA NA NA NA NA POR ......
我认为整个事情可以通过单个xmlToDataFrame调用来完成，但我对XML包不太熟悉，无法使其工作。
我真正感兴趣的是'TIME'列以及从xmlToDataFrame调用中提取的所有列。
答案 0 :(得分：14)
require(XML)
Fun1 <- function(xdata){
dum <- xmlParse(xdata)
xDf <- xmlToDataFrame(nodes = getNodeSet(dum, "//*/POR"), stringsAsFactors = FALSE)
xattrs <- xpathSApply(dum, "//*/POR/@TIME")
xDf$name <- "POR"
xDf$TIME <- xattrs
xDf
}
Fun2 <-function(xdata){
dumFun <- function(x){
xname <- xmlName(x)
xattrs <- xmlAttrs(x)
c(sapply(xmlChildren(x), xmlValue), name = xname, xattrs)
}
dum <- xmlParse(xdata)
as.data.frame(t(xpathSApply(dum, "//*/POR", dumFun)), stringsAsFactors = FALSE)
}
> identical(Fun1(xdata), Fun2(xdata))
[1] TRUE
library(rbenchmark)
benchmark(Fun1(xdata), Fun2(xdata))
test replications elapsed relative user.self sys.self user.child
1 Fun1(xdata) 100 1.047 2.069 1.044 0 0
2 Fun2(xdata) 100 0.506 1.000 0.504 0 0
sys.child
1 0
2 0
答案 1 :(得分：0)
user1609452的修改版本：
extractXML <-function(xdata, expr, transpo = T){
# expr should be "//*/Array"
# cat("[INFO] - expr is an expression of the path usually something like '//*/Array'.")
# cat("\n[INFO] - Use Transpo = F is you do not want to transpose the output. Just try it out.\n\n")
dumFun <- function(x){
xname <- xmlName(x)
xattrs <- xmlAttrs(x)
c(sapply(xmlChildren(x), xmlValue), name = xname, xattrs)
}
dum <- xmlParse(xdata)
listxml <- xpathSApply(dum, expr, dumFun)
if( transpo == T ) {
data <- as.data.table(t(listxml), stringsAsFactors = FALSE)
} else {
data <- as.data.table(rbind.fill(lapply(listxml,function(y){as.data.frame(y,stringsAsFactors=F)})))
}
return(data)
}
目标是在没有Child（transpo = F）时获取属性，但是你想获得属性。
以下示例：
<Arrays>
<Array Factor="1.000000" CompressionRate="" CompressionType="" BitsPerPixel="16" Height="515" Width="682" Name="Exp1Cam1" Type="Image"/>
<Array Factor="1.000000" CompressionRate="" CompressionType="" BitsPerPixel="16" Height="515" Width="682" Name="Exp1Cam2" Type="Image"/>
<Array Factor="1.000000" CompressionRate="" CompressionType="" BitsPerPixel="16" Height="515" Width="682" Name="Exp1Cam1" Type="Image"/>
<Array Factor="1.000000" CompressionRate="" CompressionType="" BitsPerPixel="16" Height="515" Width="682" Name="Exp1Cam2" Type="Image"/>
<Array Factor="1.000000" CompressionRate="" CompressionType="" BitsPerPixel="16" Height="515" Width="682" Name="Exp1Cam1" Type="Image"/>
<Array Factor="1.000000" CompressionRate="" CompressionType="" BitsPerPixel="16" Height="515" Width="682" Name="Exp1Cam2" Type="Image"/>
<Array Factor="1.000000" CompressionRate="" CompressionType="" BitsPerPixel="16" Height="515" Width="682" Name="Exp1Cam1" Type="Image"/>
<Array Factor="1.000000" CompressionRate="" CompressionType="" BitsPerPixel="16" Height="515" Width="682" Name="Exp1Cam2" Type="Image" Description=""/>
</Arrays>
extractXML(xdata, "//*/Array", T)
V1 V2 V3 V4 V5 V6 V7 V8
1: <list> <list> <list> <list> <list> <list> <list> <list>
extractXML(xdata, "//*/Array", F)
name Factor CompressionRate CompressionType BitsPerPixel Height Width Name Type Description
1: Array 1.000000 16 515 682 Exp1Cam1 Image NA
2: Array 1.000000 16 515 682 Exp1Cam2 Image NA
3: Array 1.000000 16 515 682 Exp1Cam1 Image NA
4: Array 1.000000 16 515 682 Exp1Cam2 Image NA
5: Array 1.000000 16 515 682 Exp1Cam1 Image NA
6: Array 1.000000 16 515 682 Exp1Cam2 Image NA
7: Array 1.000000 16 515 682 Exp1Cam1 Image NA
8: Array 1.000000 16 515 682 Exp1Cam2 Image