有没有更有效的方式拆分列

Question

There are a few values that do not import correctly when performing this read.table:

hs.industry <- read.table("https://download.bls.gov/pub/time.series/hs/hs.industry", header = TRUE, fill = TRUE, sep = "\t", quote = "", stringsAsFactors = FALSE)

Specifically there are a few values where the industry_code and industry_name are joined as a single value in the industry_code column (not sure why). Given that each industry_code is 4 digits, my approach to split and correct is:

for (i in 1:nrow(hs.industry)) {
  if (isTRUE(nchar(hs.industry$industry_code[i]) > 4)) {
    hs.industry$industry_name[i] <- gsub("[[:digit:]]","",hs.industry$industry_code[i])
    hs.industry$industry_code[i] <- gsub("[^0-9]", "",hs.industry$industry_code[i])
  }
}

I feel this is terribly innificent, but I'm not sure what approach would be better.

Thanks!

Answer 1

The problem is that lines 29 and 30 (rows 28 and 29, if we're not counting the header) have a formatting error. They use 4 spaces instead of a proper tab character. A bit of extra data cleaning is needed.

Use readLines to read in the raw text, correct the formatting error, and then read in the cleaned table:

# read in each line of the file as a list of character elements
hs.industry <- readLines('https://download.bls.gov/pub/time.series/hs/hs.industry')

# replace any instances of 4 spaces with a tab character
hs.industry <- gsub('\\W{4,}', '\t', hs.industry)

# collapse together the list, with each line separated by a return character (\n)
hs.industry <- paste(hs.industry, collapse = '\n')

# read in the new table
hs.industry <- read.table(text = hs.industry, sep = '\t', header = T, quote = '')

Answer 2

You should not have to loop through each instance, instead identify only those entries which are problematic and gsub only those entries:

replace_indx <- which(nchar(hs.industry$industry_code) > 4)
hs.industry$industry_name[replace_indx] <- gsub("\\d+\\s+", "", hs.industry$industry_code[replace_indx])
hs.industry$industry_code[replace_indx] <- gsub("\\D+", "", hs.industry$industry_code[replace_indx])

I also used "\\\\d+\\\\s+" to improve the string replacement, here I also replace the spaces:

gsub("[[:digit:]]","",hs.industry$industry_code[replace_indx])
# [1] "    Dimension stone"          "    Crushed and broken stone"

gsub("\\d+\\s+", "", hs.industry$industry_code[replace_indx])
# [1] "Dimension stone"          "Crushed and broken stone"