Ndtyjky

There's text data like this

※ 19:20 AAAA (BBB:CCC)    --DDDD aaaa         33.1   bbbb           23.8   cccc     20.9   ddddd       14.9 eeeee        8.3   ffffff        6.8   gggggg          4.5   hhhhhh 4.2    --EEEE aaaaa     11.8

or there's HTML File like this

<div style='AAAA'>

※ 19:20 AAAA (BB:CC)<br /><br /><br /><br />--DDDD<br />aaaa       33.1   bbbb         23.8   cccc   20.9<br /><br /><br />--EEEE<br />aaaaa         11.8

What I want to make in Dataframe is

Time     Type1      Type2     Data1         Data2

19:20    AAAAA      DDDD      aaaa          33.1

19:20    AAAAA      DDDD      bbbb          23.8

19:20    AAAAA      EEEE      aaaaa         11.8

How can I make it?

1) First remove everything within parens including the parens giving txt2. Then split what is left at -- separators and flatten it giving txt3. Then scan in the words separately for each element giving s and convert that to matrix m and data frame DF. Finally convert the types of the columns. No packages are used.

txt2 <- gsub("[(].*?[)]", "", txt)

txt3 <- unlist(strsplit(txt2, "--"))

s <- lapply(txt3, function(x) scan(text = x, what = "", quiet = TRUE))

m <- cbind(s[[1]][1], s[[1]][2], do.call("rbind", lapply(s[-1], 

  function(x) cbind(x[1], matrix(x[-1],,2, byrow = TRUE)))))

DF <- as.data.frame(m, stringsAsFactors = FALSE)

DF <- lapply(DF, type.convert, stringsAsFactors = FALSE)

giving:

> DF

     V1   V2   V3     V4   V5

1 19:20 AAAA DDDD   aaaa 33.1

2 19:20 AAAA DDDD   bbbb 23.8

3 19:20 AAAA DDDD   cccc 20.9

4 19:20 AAAA DDDD  ddddd 14.9

5 19:20 AAAA DDDD  eeeee  8.3

6 19:20 AAAA DDDD ffffff  6.8

7 19:20 AAAA DDDD gggggg  4.5

8 19:20 AAAA DDDD hhhhhh  4.2

9 19:20 AAAA EEEE  aaaaa 11.8

2) If we can assume that only the fifth column is numeric then we can use this simpler alternate approach. First remove the parens and everything inside them as above and then scan in the words into s. Locate the positions ix of the words containing only digits and dot and then pick out each field assembling them into a data.frame.

txt2 <- gsub("[(].*?[)]", "", txt)

s <- scan(text = txt2, what = "", quiet = TRUE)

ix <- grep("^[0-9.]+$", s)

data.frame(

   V1 = s[1], 

   V2 = s[2], 

   V3 = sub("--", "", s[sapply(ix-2, function(i) tail(grep("--", s[seq(i)]), 1))]),

   V4 = s[ix-1],

   V5 = as.numeric(s[ix]), 

   stringsAsFactors = FALSE

)

Note

The input is assumed to be:

txt <- "19:20 AAAA (BBB:CCC)    --DDDD aaaa         33.1   bbbb           23.8   cccc     20.9   ddddd       14.9 eeeee        8.3   ffffff        6.8   gggggg          4.5   hhhhhh 4.2    --EEEE aaaaa     11.8"