无法在 R 中使用 xpathSapply 选择特定的 html 表

Question

我正在尝试从以下链接http://cepea.esalq.usp.br/frango/?page=379&Dias=15 中抓取第二张表

我已经使用 XML 包尝试了以下 R 代码：

    p_frango_resfriado <- htmlTreeParse("http://cepea.esalq.usp.br/frango/?page=379&Dias=15", 
    useInternalNodes = TRUE, 
    encoding = "UTF-8")

    xpathSApply(p_frango_resfriado, "//table[@width = '95%']//tr//td[2]", xmlValue)
    xpathSApply(p_frango_resfriado, "//table[@width = '95%']//tr//td[3]", xmlValue)
    xpathSApply(p_frango_resfriado, "//table[@width = '95%']//tr//td[4]", xmlValue)

问题是这段代码抓取了网页中的两个html表，我只想抓取第二个。我试过了，下面的代码没有返回任何有趣的东西：

xpathSApply(p_frango_resfriado, 
"//a[text() = 'Preços do frango resfriado CEPEA/ESALQ - Estado SP']/table[@width = '95%']", 
xmlValue)

谁能帮我解决这个问题？我不太擅长 XPath 语言和 html。

Answer 1

在 XPath 查询中使用 XML::xmlToDataFrame

library("httr")
library("XML")
URL <- "http://cepea.esalq.usp.br/frango/?page=379&Dias=15"
temp <- tempfile(fileext = ".html")
GET(url = URL, user_agent("Mozilla/5.0"), write_disk(temp))

这两个表的唯一区别是在 xpath 查询中使用的表名

表 1：Preços do frango congelado CEPEA/ESALQ - Estado SP

xpexpr <- "//center/a[contains(., 'do frango congelado')]/../table/tr/td/font/tr"

表 2：Preços do frango resfriado CEPEA/ESALQ - Estado SP

xpexpr <- "//center/a[contains(., 'do frango resfriado')]/../table/tr/td/font/tr"

doc <- htmlParse(temp)
listofTableNodes <- getNodeSet(doc, xpexpr)
length_nodes <- length(listofTableNodes)
include_indices1 <- 1:(length_nodes - 2)

# create dataframe using xmlvalues of the nodelist. Both `getNodeSet()` 
# and `xpathSApply` will provide identical results.
# using `getNodeSet()`
df <- xmlToDataFrame(listofTableNodes[include_indices1], stringsAsFactors=FALSE)
# using `xpathSApply`
df <- xmlToDataFrame(xpathSApply(doc, xpexpr)[include_indices1], stringsAsFactors=FALSE)

# clean data
df$td <- as.Date(gsub("[Â ]\\s*", "", df$td), format = "%d/%m/%Y")
df[, 4] <- gsub("\t$", '', df[, 4])

# add column names
xpexpr <- "//center/a[contains(., 'do frango resfriado')]/../table/tr/td/font/text()"
# for Table-1
# xpexpr <- "//center/a[contains(., 'do frango congelado')]/../table/tr/td/font/text()"
listofTableNodes <- getNodeSet(doc, xpexpr)
colnames(df) <- c('Date', sapply(listofTableNodes, xmlValue))
df
#            Date Valor R$ Var./dia Var./mês
#   1  2016-08-17     4,37    0,46%     8,17%
#   2  2016-08-16     4,35    0,46%     7,67%
#   3  2016-08-15     4,33    0,46%     7,18%
#   4  2016-08-12     4,31    0,00%     6,68%
#   5  2016-08-11     4,31    0,70%     6,68%
#   6  2016-08-10     4,28    0,47%     5,94%
#   7  2016-08-09     4,26   -0,70%     5,45%
#   8  2016-08-08     4,29    3,87%     6,19%
#   9  2016-08-05     4,13    0,49%     2,23%
#   10 2016-08-04     4,11    0,00%     1,73%
#   11 2016-08-03     4,11    1,73%     1,73%
#   12 2016-08-02     4,04    0,00%     0,00%
#   13 2016-08-01     4,04    0,00%     0,00%
#   14 2016-07-29     4,04    0,00%    -0,49%
#   15 2016-07-28     4,04   -0,25%    -0,49%

注意：每天，此网页上的值都会更新，将使用length_nodes 将其考虑在内。

在没有 XPath 查询的情况下使用 XML::readHTMLTable

library("httr")
library("XML")
URL <- "http://cepea.esalq.usp.br/frango/?page=379&Dias=15"
temp <- tempfile(fileext = ".html")
GET(url = URL, user_agent("Mozilla/5.0"), write_disk(temp))
df <- readHTMLTable(temp, stringAsFactors = FALSE, which = 8)
# Table 1
df[4:18,]
# Table 2
df[28:42,]

在 XPath 查询中使用 XML::readHTMLTable

library("httr")
library("XML")
URL <- "http://cepea.esalq.usp.br/frango/?page=379&Dias=15"
temp <- tempfile(fileext = ".html")
GET(url = URL, user_agent("Mozilla/5.0"), write_disk(temp))
doc <- htmlParse(temp)

# XPath Query
# Table -1
xpexpr <- "//center/a[contains(., 'do frango congelado')]/../table/tr/td/font"
df <- xpathSApply(doc, xpexpr, readHTMLTable)
include_indices <- 1:(nrow(df[[4]]) -4)
df <- df[[4]][include_indices,]

# Table-2
xpexpr <- "//center/a[contains(., 'do frango resfriado')]/../table/tr/td/font"
df <- xpathSApply(doc, xpexpr, readHTMLTable)
include_indices <- 1:(nrow(df[[4]]) -4)
df <- df[[4]][include_indices,]

Answer 2

这现在应该可以工作，但我想知道如果你每天运行它是否会正常工作。

options(stringsAsFactors=FALSE)
library(XML)

#get website
tfile <- tempfile()
download.file("http://cepea.esalq.usp.br/frango/?page=379&Dias=15", tfile)
temp <- readHTMLTable(tfile)

#read in table
tbl <- temp[[5]][!is.na(temp[[5]]$V2) & !is.na(temp[[5]]$V4),]
tbl2 <- split(tbl, cumsum(tbl$V1==""))[[2]][-1,]
tbl2 

#perform formatting
colnames(tbl2) <- c("Date","Price","Pct1","Pct2")
tbl2 <- data.frame(apply(tbl2, c(1,2), function(x) {
    x <- gsub("Â","",x)
    x <- gsub(",",".",x,fixed=TRUE)
    x <- gsub("%","",x)
}))
tbl2$Date <- as.Date(tbl2$Date,"  %d/%m/%Y  ")
tbl2$Price <- as.numeric(tbl2$Price)
tbl2$Pct1 <- as.numeric(tbl2$Pct1) / 100
tbl2$Pct2 <- as.numeric(tbl2$Pct2) / 100
tbl2

#         Date Price    Pct1    Pct2
#65 2016-08-17  4.37  0.0046  0.0817
#66 2016-08-16  4.35  0.0046  0.0767
#67 2016-08-15  4.33  0.0046  0.0718
#68 2016-08-12  4.31  0.0000  0.0668
#69 2016-08-11  4.31  0.0070  0.0668
#70 2016-08-10  4.28  0.0047  0.0594
#71 2016-08-09  4.26 -0.0070  0.0545
#72 2016-08-08  4.29  0.0387  0.0619
#73 2016-08-05  4.13  0.0049  0.0223
#74 2016-08-04  4.11  0.0000  0.0173
#75 2016-08-03  4.11  0.0173  0.0173
#76 2016-08-02  4.04  0.0000  0.0000
#77 2016-08-01  4.04  0.0000  0.0000
#78 2016-07-29  4.04  0.0000 -0.0049
#79 2016-07-28  4.04 -0.0025 -0.0049