fread fails if whitespace before first character

[dpastoor]

The program I use by default gives output like so:

" ITERATION    THETA1       THETA2"                
"            0  3.95527E+01  2.10651E+01"

I first tried with the stable version (1.9.4) and got the following error:

Expected sep (',') but new line, EOF (or other non printing character) ends field 1 on line 4 when detecting types:             0  3.95527E+01  2.10651E+01 

with the development version (1.9.5) I get the following error:

Not positioned correctly after testing format of header row. ch=' '

This seems like it would be resolved by #758 or #558 (variable whitespace delimiter)

but just wanted to give another example of a situation where fread currently fails and need to fall back to read.table(sep="") to get the table to read in properly

[GHarmata]

I am experiencing a similar problem; however, the R session completely aborts in 1.9.4, instead of giving a warning message.

Here is an example illustrating the format of dataset with which I am working:

"  22 4 6 4" 
"  34 22 34 5"
"  6 2 1 4"

Each row begins with two white spaces, but the separator for the rest of the values is a single white space.

In version 1.9.4, when I attempted to use fread() in RStudio with my much-larger dataset, I received the following pop-up warning (and the session had to be restarted):

R Session Aborted
R encountered a fatal error. The session was terminated.

---

Using version 1.9.5, I instead received this error message:

Not positioned correctly after testing format of header row. ch=' '

I found this output very confusing, since I had specified that there was no header row (header = FALSE). The verbose=TRUE version of this output is below.

Input contains no \n. Taking this to be a filename to open
File opened, filesize is 0.024641 GB.
Memory mapping ... ok
Detected eol as \r\n (CRLF) in that order, the Windows standard.
Positioned on line 1 after skip or autostart
This line is the autostart and not blank so searching up for the last non-blank ... line 1
Detecting sep ... ' '
Detected 561 columns. Longest stretch was from line 1 to line 30
Starting data input on line 1 (either column names or first row of data). First 10 characters:   2.571777
Error in fread("UCI HAR Dataset/test/X_test.txt", header = FALSE, verbose = TRUE) : 
  Not positioned correctly after testing format of header row. ch=' '

I am using R version 3.20 ("Full of Ingredients"), with RStudio version 0.98.1103.

[jaapwalhout]

I had the same problem. This works:

 df <- read.table("./data/train/X_train.txt")

but reading the same file with fread does not work and throws me an error:

> dt <- fread("./data/train/X_train.txt")
Error in fread("./data/train/X_train.txt") : 
  Not positioned correctly after testing format of header row. ch=' '

The file in question can be downloaded from here (66MB's).

[arunsrinivasan]

Fixed in devel. Please upgrade and test.

[jaapwalhout]

@arunsrinivasan Tested. It now works on the file I specified above.