Data.table: fread may fail to parse valid file when dec=','

Created on 13 Apr 2018 · 3Comments · Source: Rdatatable/data.table

> fread('a,b,c,d\n1e1,1e2,1e3,"4,0001"\n1,2,3,4\n', dec=',')
       a     b     c      d
   <num> <num> <num>  <num>
1:    10   100  1000 4.0001
Warning message:
In fread("a,b,c,d\n1e1,1e2,1e3,\"4,0001\"\n1,2,3,4\n", dec = ",") :
  Discarded single-line footer: <<1,2,3,4>>

This happens because float parser greedily consumes 1,2 as a single token, whereas without quotes it must be parsed as 2 separate fields.

In addition, the "Details" section in documentation has the following information (which is long since being outdated):

‘fread’ uses C function ‘strtod’ to read numeric data; e.g.,
‘1.23’ or ‘1,23’. ‘strtod’ retrieves the decimal separator (‘.’ or
‘,’ usually) from the locale of the R session rather than as an
argument passed to the ‘strtod’ function. So for
‘fread(...,dec=",")’ to work, ‘fread’ changes this (and only this)
R session's locale temporarily to a locale which provides the
desired decimal separator.

bug fread

Source

st-pasha

👍1

All 3 comments

Awesome package! Thank you.
Coincidentally, I experienced a possibly related strange behaviour with fread yesterday.
Here is a simple example (derived from a real case using lapply fread on a bunch of files).

library(data.table)

DT = data.table(A = rep("20,1", 1e4))
fwrite(DT, "DT.csv", quote = FALSE)

classA = character(1e3)

for (i in seq_along(classA)) {

  DT = fread("DT.csv", sep = ";", dec = ",", colClasses = "numeric")
  classA[i] = DT[, class(A)]
}

table(classA)

This gives me:

Warning message:
In fread("DT.csv", sep = ";", dec = ",", colClasses = "numeric",  :
  Bumped column 1 to type character on data row 387, field contains '20,1'. Coercing previously read values in this column from logical, integer or numeric back to character which may not be lossless; e.g., if '00' and '000' occurred before they will now be just '0', and there may be inconsistencies with treatment of ',,' and ',NA,' too (if they occurred in this column before the bump). If this matters please rerun and set 'colClasses' to 'character' for this column. Please note that column type detection uses a sample of 1,000 rows (100 rows at 10 points) so hopefully this message should be very rare. If reporting to datatable-help, please rerun and include the output from verbose=TRUE.

table(classA)
# classA
# character   numeric 
#        1       999 
which(classA == "character")
# [1] 9
`````

So, sometimes, the column is read as character with the spotted row index being different for different runs. And the ```which``` indicates it is more likely to happen in the first iterations. I don't get the randomness and the fact that colClasses is ignored (I set colClasses after reading fread doc...).
And... I did not manage to reproduce the error when setting verbose = TRUE... Also, the bug was observed using RStudio, not reproduced in the R console...
Sorry if I missed or misunderstood something...

```r
sessionInfo()
R version 3.3.2 (2016-10-31)
Platform: x86_64-w64-mingw32/x64 (64-bit)
Running under: Windows 7 x64 (build 7601) Service Pack 1

locale:
[1] LC_COLLATE=English_United States.1252  LC_CTYPE=English_United States.1252   
[3] LC_MONETARY=English_United States.1252 LC_NUMERIC=C                          
[5] LC_TIME=English_United States.1252    

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base     

other attached packages:
[1] data.table_1.10.4-3

loaded via a namespace (and not attached):
[1] tools_3.3.2 yaml_2.1.18

Atrebas on 14 Apr 2018

please make sure to update to the development version of data.table and test again. there have been a lot of improvements to fread since the last release