从R中的文本文件中提取列表数据

Question

There is probably already a post on this topic, but I'm not sure what terms to search. I am trying to import data from a txt file with this format (the first 2 lines of which are not of interest):

FN Clarivate Analytics Web of Science
VR 1.0
PT J
AU Ahituv, Nadav
   Zhu, Yiwen
   Visel, Axel
   Holt, Amy
   Afzal, Veena
   Pennacchio, Len A.
   Rubin, Edward M.
TI Deletion of ultraconserved elements yields viable mice
SO PLOS BIOLOGY
VL 5
IS 9
BP 1906
EP 1911
AR e234
DI 10.1371/journal.pbio.0050234
PD SEP 2007
PY 2007
RI Visel, Axel/A-9398-2009; Ahituv, Nadav/; Pennacchio, Len/
OI Visel, Axel/0000-0002-4130-7784; Ahituv, Nadav/0000-0002-7434-8144;
   Pennacchio, Len/0000-0002-8748-3732
SN 1544-9173
UT WOS:000249552300010
PM 17803355
ER

PT J
AU Ahmadiyeh, Nasim
   Pomerantz, Mark M.
   Grisanzio, Chiara
   Herman, Paula
   Jia, Li
   Almendro, Vanessa
   He, Housheng Hansen
   Brown, Myles
   Liu, X. Shirley
   Davis, Matt
   Caswell, Jennifer L.
   Beckwith, Christine A.
   Hills, Adam
   MacConaill, Laura
   Coetzee, Gerhard A.
   Regan, Meredith M.
   Freedman, Matthew L.
TI 8q24 prostate, breast, and colon cancer risk loci show tissue-specific
   long-range interaction with MYC
SO PROCEEDINGS OF THE NATIONAL ACADEMY OF SCIENCES OF THE UNITED STATES OF
   AMERICA
VL 107
IS 21
BP 9742
EP 9746
DI 10.1073/pnas.0910668107
PD MAY 25 2010
PY 2010
RI Davis, Matt/F-9045-2012; He, Housheng/G-9614-2011; he, housheng hansen/; Caswell-Jin, Jennifer/; Brown, Myles/
OI he, housheng hansen/0000-0003-2898-3363; Caswell-Jin,
   Jennifer/0000-0002-5711-8355; Brown, Myles/0000-0002-8213-1658
SN 0027-8424
UT WOS:000278054700049
PM 20453196
ER

Since some of the categories (eg AU) have more than one object, I think I need to import as a list. The category labels are all 2 characters followed by a space, but some categories are on more than one line, and subsequent lines are not labeled with the category label. In addition, for some categories that take up more than one line, such as AU, I would like the data to be imported as a vector. For others, suce as TI or SO, I would like to catenate the multiple lines into one object of class character in the list.

I would like the entries to look something like this:

print(<portion of list that corresponds to AU for first reference>)
[AU]
[[1]] "Ahituv, Nadav"      "Zhu, Yiwen"         "Visel, Axel"        "Holt, Amy"          "Afzal, Veena"      
[[6]] "Pennacchio, Len A." "Rubin, Edward M."

print(<portion of lilst that corresponds to TI and SO for second reference>)
[TI]
[[1]] "8q24 prostate, breast, and colon cancer risk loci show tissue-specific long-range interaction with MYC"
[SO]
[[1]] "PROCEEDINGS OF THE NATIONAL ACADEMY OF SCIENCES OF THE UNITED STATES OF AMERICA"

I've tried using scan() using the following code:

scan("savedrecs_spitz refs.txt", what = "character", sep = "\n")

However, what gets read in is a single character vector where each line of the txt is read in as a separate object in the vector:

[1] "FN Clarivate Analytics Web of Science" "VR 1.0"                                  
[3] "PT J"                                     "AU Ahituv, Nadav"                        
[5] "   Zhu, Yiwen"                            "   Visel, Axel"

Is there a different function I should be using to read in these data?

Answer 1

Is that what you're looking for?

dt=scan("savedrecs_spitz refs.txt", what = "character", sep = "\n")

mgrep=function(dt){
  v=intersect(grep("[A-Z]{2}",dt),which(nchar(dt)==2))
  ret=list()
  for(i in 1:length(v)){
    end=ifelse(i==length(v),length(dt),(v[i+1]-1))
    st=(v[i]+1)
    ret[[i]]=dt[st:end]
  }
  names(ret)=dt[v]
  return(ret)
}
mgrep(dt)

ps: pay attention to special characters being wrongly read like "FN" that would not be properly used inside the function.

Answer 2

Think I have solved your problem, but I save the data in a data.frame

library(stringr)

text <- scan("text.txt",sep = "\n",what = "character")

textLoop <- grep("^[[:upper:]]|^[[:blank:]]", text, value = TRUE)

for(i in 1:length(textLoop)){
  if(grepl("^[[:blank:]]", textLoop[i])){
    partOne <- substring(textLoop[i-1], 1, 2)
    textLoop[i] <- paste0(partOne, textLoop[i])
  }
}

textDf <- data.frame(partOne = substring(textLoop, 1, 2),
                     partTwo = substring(textLoop, 4))