CODEWITHSUNDEEP
regexxmlrxml-parsing
Andrew Brown
Andrew Brown

Reputation: 333

How to use regular expressions in XML to extract a string from a list

I have just begun using the XML package in R and I'm having trouble extracting a string from xml list:

> library("XML")
> library("stringr")
> url = "html-1.html"
> parsed_doc = htmlParse(file=url, useInternalNodes = TRUE)
> products <- getNodeSet(doc = parsed_doc, path = "//li[contains(.,Product ID')]")
> products
[[1]]
<li>&#13;
                        Product ID:&#13;
                        000002434482&#13;
                    </li> 

[[2]]
<li>&#13;
                        Product ID:&#13;
                        000002183105&#13;
                    </li> 

[[3]]
<li>&#13;
                        Product ID:&#13;
                        000002183105&#13;
                    </li>

I would like to create a vector containing each ID. I have tried a number of regularized expressions to extract the 12 digit product ID but can't seem to get it to work.

> mrn <- str_extract(products , "[[:digit:]{12}")
> mrn <- str_extract(products , "[[:digit:]+&#13]")
> mrn <- str_extract(products , "[0-9]+&#13")

I wondered if the list structure had something to do with it or maybe the spacing?

I have also tried > mrn <- str_extract(products , ".{16}") however, R returns pointer values such as "<pointer: 0x6815" - I think this is close but I'm not sure what this means.

Upvotes: 1

Views: 1449

Answers (1)

Wiktor Stribiżew
Wiktor Stribiżew

Reputation: 627022

You are almost there. The node set is not a string vector, you need to get the string values out of it first. You can easily extract them with xmlValue, and then you can use str_extract (or str_extract_all):

> v <- sapply(products, xmlValue)
> v
[1] "\r\n                        Product ID:\r\n                        000002434482\r\n"                    
[2] "\r\n                        Product ID:\r\n                        000002183105\r\n"                    
[3] "\r\n                        Product ID:\r\n                        000002183105\r\n                    "
> unlist(str_extract_all(v, "[[:digit:]]+"))
[1] "000002434482" "000002183105" "000002183105"

If the IDs are whole words containing 12 digits only, you can use a more precise expression like

"\\b[[:digit:]]{12}\\b"

where \b is a word boundary and {12} is a limiting quantifier matching exactly 12 occurrences of a digit.

Alternatively, you can also extract these IDs with str_match and Product ID:\s*(\d{12})\b regex that matches Product ID: + space(s) + the 12-digit whole word number (that is captured, thus, we need to use str_match rather than str_extract):

> res <- unlist(str_match(v, "Product ID:\\s*(\\d{12})\\b"))
> res[,2]
[1] "000002434482" "000002183105" "000002183105"

Upvotes: 1

Related Questions