读取R中的多个xml文件并合并数据

时间:2018-03-09 15:08:47

标签: r xml for-loop xbrl

我有一个包含超过1000个带扩展名的文件的文件夹(但它们不是真正的 xml文件)。

我想自动从这些文件中提取某些内容,以便矩阵或表格是最终结果(我可以在R中进一步使用它进行分析,或导出到1 csv文件等)。

我已经制作/更改了适用于单个文件的代码,但是无法让它自动完成其余的工作。循环?

所以我的单个文件的代码如下:

library(xml2)

temp <- read_xml("test.xml")
# get all the <ns2:opendataField>s
recs <- xml_find_all(temp, "//ns2:opendataField")
# extract and clean all the columns
vals <- trimws(xml_text(recs))
#create columns
cols <- xml_attr(xml_find_all(temp, "//ns2:opendataField"), "key")
#create rows
rows <- xml_attr(xml_find_all(temp, "//ns2:opendataField"), "value")
datakvk <- data.frame(cols,rows)

这导致:

 > head(datakvk)
                                              cols       rows
1                                  SbiBusinessCode      18129
2                             DocumentAdoptionDate 2017-08-22
3                                    FinancialYear       2016
4                                     BalanceSheet       <NA>
5 BalanceSheetBeforeAfterAppropriationResultsTitle       <NA>
6      BalanceSheetBeforeAfterAppropriationResults         Na
> 

最后,对于所有这1000个文件,我希望得到类似的东西:

                                              cols       file 1   file 2
1                                  SbiBusinessCode      18129     34234
2                             DocumentAdoptionDate 2017-08-22     452454
3                                    FinancialYear       2016     2016
4                                     BalanceSheet       <NA>     2016
5 BalanceSheetBeforeAfterAppropriationResultsTitle       <NA>     <NA>
6      BalanceSheetBeforeAfterAppropriationResults         Na
> 

我尝试了以下代码,但它没有用:

list.files(pattern=".xml$") #

# create a list from these files
list.filenames<-list.files(pattern=".xml$")

# create an empty list that will serve as a container to receive the incoming files
list.data<-list()

# create a loop to read in your data
for (i in 1:length(list.filenames))
{
  list.data[[i]]<-read_xml(list.filenames[i])
  recs <- xml_find_all(list.data[[i]], "//ns2:opendataField")
  vals <- trimws(xml_text(recs))
  cols <- xml_attr(xml_find_all(list.data[[i]], "//ns2:opendataField"), "value")
  rows <- xml_attr(xml_find_all(list.data[[i]], "//ns2:opendataField"), "key")
}

# add the names of  data to the list
names(list.data)<-list.filenames

我错过了什么?我哪里出错了?

先谢谢你帮助我....

完成:(一个单一的源文件(1000个中的一个看起来像:):

<?xml version="1.0" encoding="UTF-8" standalone="yes"?>
<opendata xmlns:ns2="http://schemas.kvk.nl/xb/query/service/2016/1/0/0">
  <ns2:opendataField key="SbiBusinessCode" value="18129"/>
  <ns2:opendataField key="DocumentAdoptionDate" value="2017-08-22"/>
  <ns2:opendataField key="FinancialYear" value="2016"/>
  <ns2:opendataField key="BalanceSheet">
    <ns2:opendataField key="BalanceSheetBeforeAfterAppropriationResultsTitle">
      <ns2:opendataField key="BalanceSheetBeforeAfterAppropriationResults" value="Na"/>
    </ns2:opendataField>
    <ns2:opendataField key="BalanceSheetTitle">
      <ns2:opendataField key="Assets" value="61296">
        <ns2:opendataField key="AssetsNoncurrent" value="8978">
          <ns2:opendataField key="IntangibleAssets" value="8978"/>
        </ns2:opendataField>
        <ns2:opendataField key="AssetsCurrent" value="52318">
          <ns2:opendataField key="Inventories" value="2239"/>
          <ns2:opendataField key="Receivables" value="40560"/>
          <ns2:opendataField key="CashAndCashEquivalents" value="9519"/>
        </ns2:opendataField>
      </ns2:opendataField>
      <ns2:opendataField key="EquityAndLiabilities" value="61296">
        <ns2:opendataField key="Equity" value="201">
          <ns2:opendataField key="ShareCapital" value="1"/>
          <ns2:opendataField key="ReservesOther" value="200"/>
        </ns2:opendataField>
        <ns2:opendataField key="LiabilitiesCurrent" value="61095"/>
      </ns2:opendataField>
    </ns2:opendataField>
  </ns2:opendataField>
</opendata>

2 个答案:

答案 0 :(得分:0)

您可以使用lapply,然后使用cbind do.call library(xml2) library(dplyr) files <- list.files(pattern = ".xml$") data <- lapply(files, function(x) { temp <- read_xml(x) %>% xml_find_all("//ns2:opendataField") cols <- xml_attr(xml_find_all(temp, "//ns2:opendataField"), "key") rows <- xml_attr(xml_find_all(temp, "//ns2:opendataField"), "value") out <- data.frame(rows, row.names = cols) names(out) <- x out }) do.call(cbind, data) 列表元素:

                                                  file1.xml  file2.xml
SbiBusinessCode                                       18129      18129
DocumentAdoptionDate                             2017-08-22 2017-08-22
FinancialYear                                          2016       2016
BalanceSheet                                           <NA>       <NA>
BalanceSheetBeforeAfterAppropriationResultsTitle       <NA>       <NA>
BalanceSheetBeforeAfterAppropriationResults              Na         Na
BalanceSheetTitle                                      <NA>       <NA>
Assets                                                61296      61296
AssetsNoncurrent                                       8978       8978
IntangibleAssets                                       8978       8978
AssetsCurrent                                         52318      52318
Inventories                                            2239       2239
Receivables                                           40560      40560
CashAndCashEquivalents                                 9519       9519
EquityAndLiabilities                                  61296      61296
Equity                                                  201        201
ShareCapital                                              1          1
ReservesOther                                           200        200
LiabilitiesCurrent                                    61095      61095

输出具有相同内容的两个文件:

{{1}}

答案 1 :(得分:0)

考虑将您的for循环转换为调用lapply data.frame()的{​​{1}}以获取数据框列表。并且因为您的XML文件可能具有不同的键/值,所以关闭数据帧列表的简单cbind将不起作用,因此请使用与Reduce()的链合并,保留所有行(即,完全外部联接) )。

...
# BUILD DATAFRAME LIST
df_list <- lapply(list.filenames, function(f) {
  doc <- read_xml(f)

  setNames(data.frame(
    xml_attr(xml_find_all(doc, "//ns2:opendataField"), "key"),
    xml_attr(xml_find_all(doc, "//ns2:opendataField"), "value")
  ), c("key", f))

})

# CHAIN MERGE INTO MASTER DATAFRAME
final_df <- Reduce(function(x,y) merge(x, y, by="key", all=TRUE), df_list)