`fread`-ing文本文件，带有不规则的分隔符

Question

I need to read a large ( > 1GB) text file with irregular separators. A simplified example is the following:

1,1,1
1,1,1
...
1,1,1,1
1,1,1
...

The irregular lines (3 separators instead of 2 in the above example) show up possibly after the thousandth line.

read.table works with fill=TRUE and 4 values in col.names . fread with fill=TRUE fails to read the irregular lines. Adding col.names causes an error.

A MWE:

sometext  <-  "1,1,1"
fulltext  <-  paste(rep(sometext, 150), collapse="\n")
fulltext  <-  paste0(fulltext, "\n1,1,1,1") # Irregular line
fulltext  <-  paste0(fulltext, "\n1,1,1")

# Works
mwe.readtable  <- read.table(text=fulltext,
                            sep=",",
                            col.names = paste("V", 1:4, sep=""),
                            fill=TRUE)
# Fails w/ warning
mwe.fread  <- fread(fulltext,
                   sep =  ",",
                   fill=TRUE)

# Fails w/ error
mwe.fread2  <- fread(fulltext,
                   sep =  ",",
                   col.names = paste("V", 1:4, sep=""),
                   fill=TRUE)

Is there a way to prevent data.table from trying to guess the maximum number of separators?