CODEWITHSUNDEEP
r
Autumn
Autumn

Reputation: 585

Extract multiple types of pattern from string

I am extracting multiple types of pattern from a string. For example,

"Listed 03/25/2013 for 25000 and sold for $10,250 on 4/5/2010"

I would like to extract dates "03/25/2013" "4/5/2010" to vector 'dates', and "25000" "$10,250" to vector amounts.

text <- "Listed 03/25/2013 for 25000 and sold for $10,250 on 4/5/2010"
  # extract dates
dates <- str_extract_all(text,"\\d{1,2}\\/\\d{1,2}\\/\\d{4}")[[1]]
  # extract amounts
text2 <- as.character(gsub("\\d{1,2}\\/\\d{1,2}\\/\\d{4}", " ", text))
amountsdollar <- as.character(str_extract_all(text2,"\\$\\(?[0-9,.]+\\)?"))
text3 <- as.character(gsub("\\$\\(?[0-9,.]+\\)?", " ", text2))
amountsnum <- as.character(str_extract_all(text3,"\\(?[0-9,.]+\\)?"))
amounts <- as.vector(c(amountsdollar, amountsnum))
list(dates, amounts)

But the order is not kept. Is there a better way to do it? Thanks.

Upvotes: 0

Views: 645

Answers (1)

Matthew Plourde
Matthew Plourde

Reputation: 44614

base R handles this fine

x <- "Listed 03/25/2013 for 25000 and sold for $10,250, on 4/5/2010"
date.pat <- '\\d{1,2}/\\d{1,2}/\\d{2,4}'
amount.pat <- '(?<=^| )[$,0-9]+[0-9](?=,|\\.|$| )'

dates <- regmatches(x, gregexpr(date.pat, x))
amounts <- regmatches(x, gregexpr(amount.pat, x, perl=TRUE))

Upvotes: 6

Related Questions