fread only allows URLs on `input`, not `file`

[pnacht]

The documentation for fread's argument file states

[...] a URL starting http:// [...]

However, if I try reading a file via HTTP, I can only get it to work by passing it to the generic input, not file.

library(data.table)
packageVersion("data.table")
#> [1] ‘1.14.0’

x <- "http://dados.cvm.gov.br/dados/FI/DOC/INF_DIARIO/DADOS/inf_diario_fi_202012.csv"

fread(x, sep = ";")
#>         TP_FUNDO         CNPJ_FUNDO  DT_COMPTC VL_TOTAL  VL_QUOTA VL_PATRIM_LIQ
#>      1:       FI 00.017.024/0001-53 2020-12-01  1099408 27.493767       1098370
#>      2:       FI 00.017.024/0001-53 2020-12-02  1099497 27.493801       1098371
#>      [...]

fread(file = x, sep = ";")
#> Error in fread(file = x, sep = ";"): File 'http://dados.cvm.gov.br/dados/FI/DOC/INF_DIARIO/DADOS/inf_diario_fi_202012.csv' does not exist or is non-readable. getwd()=='[...]'

^{Created on 2021-04-12 by the reprex package (v1.0.0)}

# Output of sessionInfo()

sessionInfo()
#> R version 4.0.5 (2021-03-31)
#> Platform: x86_64-w64-mingw32/x64 (64-bit)
#> Running under: Windows 10 x64 (build 19041)
#> 
#> Matrix products: default
#> 
#> locale:
#> [1] LC_COLLATE=English_United States.1252 
#> [2] LC_CTYPE=English_United States.1252   
#> [3] LC_MONETARY=English_United States.1252
#> [4] LC_NUMERIC=C                          
#> [5] LC_TIME=English_United States.1252    
#> 
#> attached base packages:
#> [1] stats     graphics  grDevices utils     datasets  methods   base     
#> 
#> loaded via a namespace (and not attached):
#>  [1] fansi_0.4.2       crayon_1.4.1      utf8_1.2.1        digest_0.6.27    
#>  [5] backports_1.2.1   lifecycle_1.0.0   reprex_1.0.0      magrittr_2.0.1   
#>  [9] evaluate_0.14     pillar_1.5.1      highr_0.8         stringi_1.5.3    
#> [13] rlang_0.4.10      fs_1.5.0          vctrs_0.3.7       ellipsis_0.3.1   
#> [17] rmarkdown_2.7     styler_1.3.2      tools_4.0.5       stringr_1.4.0    
#> [21] glue_1.4.2        purrr_0.3.4       xfun_0.22         yaml_2.2.1       
#> [25] compiler_4.0.5    pkgconfig_2.0.3   htmltools_0.5.1.1 knitr_1.31       
#> [29] tibble_3.1.0

[avimallu]

The problem lies in the conditional branch for the input argument in fread.R:

data.table/R/fread.R

Lines 50 to 117 in ec1259a

	if (input=="" || length(grep('\\n|\\r', input))) {
	# input is data itself containing at least one \n or \r
	} else {
	if (substring(input,1L,1L)==" ") {
	stop("input= contains no \\n or \\r, but starts with a space. Please remove the leading space, or use text=, file= or cmd=")
	}
	str6 = substring(input,1L,6L) # avoid grepl() for #2531
	str7 = substring(input,1L,7L)
	str8 = substring(input,1L,8L)
	if (str7=="ftps://" || str8=="https://") {
	# nocov start
	if (!requireNamespace("curl", quietly = TRUE))
	stop("Input URL requires https:// connection for which fread() requires 'curl' package which cannot be found. Please install 'curl' using 'install.packages('curl')'.") # nocov
	tmpFile = tempfile(fileext = paste0(".",tools::file_ext(input)), tmpdir=tmpdir) # retain .gz extension in temp filename so it knows to be decompressed further below
	curl::curl_download(input, tmpFile, mode="wb", quiet = !showProgress)
	file = tmpFile
	on.exit(unlink(tmpFile), add=TRUE)
	# nocov end
	}
	else if (str6=="ftp://" || str7== "http://" || str7=="file://") {
	# nocov start
	method = if (str7=="file://") "internal" else getOption("download.file.method", default="auto")
	# force "auto" when file:// to ensure we don't use an invalid option (e.g. wget), #1668
	tmpFile = tempfile(fileext = paste0(".",tools::file_ext(input)), tmpdir=tmpdir)
	download.file(input, tmpFile, method=method, mode="wb", quiet=!showProgress)
	# In text mode on Windows-only, R doubles up \r to make \r\r\n line endings. mode="wb" avoids that. See ?connections:"CRLF"
	file = tmpFile
	on.exit(unlink(tmpFile), add=TRUE)
	# nocov end
	}
	else if (length(grep(' ', input, fixed = TRUE)) && !file.exists(input)) { # file name or path containing spaces is not a command
	cmd = input
	if (input_has_vars && getOption("datatable.fread.input.cmd.message", TRUE)) {
	message("Taking input= as a system command ('",cmd,"') and a variable has been used in the expression passed to `input=`. Please use fread(cmd=...). There is a security concern if you are creating an app, and the app could have a malicious user, and the app is not running in a secure environment; e.g. the app is running as root. Please read item 5 in the NEWS file for v1.11.6 for more information and for the option to suppress this message.")
	}
	}
	else {
	file = input # filename
	}
	}
	}
	if (!is.null(cmd)) {
	(if (.Platform$OS.type == "unix") system else shell)(paste0('(', cmd, ') > ', tmpFile<-tempfile(tmpdir=tmpdir)))
	file = tmpFile
	on.exit(unlink(tmpFile), add=TRUE)
	}
	if (!is.null(file)) {
	file_info = file.info(file)
	if (is.na(file_info$size)) stop("File '",file,"' does not exist or is non-readable. getwd()=='", getwd(), "'")
	if (isTRUE(file_info$isdir)) stop("File '",file,"' is a directory. Not yet implemented.") # dir.exists() requires R v3.2+, #989
	if (!file_info$size) {
	warning("File '", file, "' has size 0. Returning a NULL ",
	if (data.table) 'data.table' else 'data.frame', ".")
	return(if (data.table) data.table(NULL) else data.frame(NULL))
	}
	ext2 = substring(file, nchar(file)-2L, nchar(file)) # last 3 characters ".gz"
	ext3 = substring(file, nchar(file)-3L, nchar(file)) # last 4 characters ".bz2"
	if (ext2==".gz" || ext3==".bz2") {
	if (!requireNamespace("R.utils", quietly = TRUE))
	stop("To read gz and bz2 files directly, fread() requires 'R.utils' package which cannot be found. Please install 'R.utils' using 'install.packages('R.utils')'.") # nocov
	FUN = if (ext2==".gz") gzfile else bzfile
	R.utils::decompressFile(file, decompFile<-tempfile(tmpdir=tmpdir), ext=NULL, FUN=FUN, remove=FALSE) # ext is not used by decompressFile when destname is supplied, but isn't optional
	file = decompFile # don't use 'tmpFile' symbol again, as tmpFile might be the http://domain.org/file.csv.gz download
	on.exit(unlink(decompFile), add=TRUE)
	}
	file = enc2native(file) # CfreadR cannot handle UTF-8 if that is not the native encoding, see #3078.
	input = file

It looks like the pattern matching using substring for URLs (https://, ftp://, http:// etc.) is performed only for the input argument and not done for file.

Judging by the code, I think the intent was to have input cater to URLs, and not have it handled in the file argument, and is perhaps an oversight in the documentation.

To others: should I file a pull request to correct the documentation reference? That's something within my capability.

[pnacht]

I noticed that in the source code, but I'm not sure if that is the actual intent or if it was an oversight when writing the code. After all, as I understand it, input is simply meant to be a delegator between text and file, with no special features of its own.

To be defined by those in the know, I suppose.

[mattdowle]

Great spot! I agree it was an oversight and file= should accept URLs; it's a bug.