Question

I have a text file with few first few lines not required and then there is a table which is like this

-Hyphen line-
| col1 | col2 | col3 col4 col5 |
-Hyphen line-
| 1 | 2:24:21 PM 3/22/2012 | 0 0 1 |
| 2 | 2:24:21 PM 3/22/2012 | 1 · 0 |

1. Col1, Col2 are separated by | but col3,col4 and col5 are just separated by space.
2. Data Type should be maintained like col2 date and col3,4,5 as number ?
3. Row 2, Col4 is dot and so it should be read ad NA
4. Hyphen lines starts and end with – – –

Question:
1. I can use scan, but how to avoid reading “|” and “-” ?
2. I can skip top few lines but how to skip say 50th lines in addition to top few lines.

Answer 1

You can read it as a table as-is, then split up the column and recombine.

txt <- "| col1 | col2 | col3 col4 col5 |
| 1 | 2:24:21 PM 3/22/2012 | 0 0 1 |
| 2 | 2:24:21 PM 3/22/2012 | 1 · 0 |"

x <- read.table(text = txt, sep = "|", header = TRUE, stringsAsFactors = FALSE)

## drop unnecessary columns from the original sep split
x <- x[,-c(1,ncol(x))]

## split the desired column by the spaces, result is a character matrix
## including an unnecessary first column
split.col3 <- do.call("rbind", strsplit(x[,3], " "))

## bind to the original, dropping the unneeded columns
cbind(x[,-3], split.col3[,-1])
  col1                   col2 1 2 3
1    1  2:24:21 PM 3/22/2012  0 0 1
2    2  2:24:21 PM 3/22/2012  1 · 0

I avoided referring to the original column names, since you said you want to skip those lines. Simply add header = FALSE and skip = 50 to the read.table call, then add whatever column names make sense afterwards.

Also, you can then strip the “.” from columns where necessary, convert to date-time formats or numbers as required. Use colClasses in read.table if you know them up front. It makes sense to me to break this down into a number of steps, rather than trying to do it all with one read function.