[ARROW-8615] [R] Error better and insist on RandomAccessFile in read_feather

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 0.16.0, 0.17.0
• Fix Version/s: 1.0.0
• Component/s: R
• Labels: pull-request-available

Environment

R version 3.6.3 (2020-02-29)
Platform: x86_64-pc-linux-gnu (64-bit)
Running under: Ubuntu 18.04.4 LTS

other attached packages:
[1] arrow_0.17.0

loaded via a namespace (and not attached):
 [1] tidyselect_1.0.0 bit_1.1-15.2 compiler_3.6.3 magrittr_1.5
 [5] assertthat_0.2.1 R6_2.4.1 glue_1.4.0 Rcpp_1.0.4.6
 [9] bit64_0.9-7 vctrs_0.2.4 rlang_0.4.5 purrr_0.3.4

• External issue URL: https://github.com/apache/arrow/issues/24778
• Assignee: Neal Richardson
• Reporter: Lucas Nesi
• Created: 28/Apr/20 15:27
• Updated: 11/Jan/23 08:01
• Resolved: 15/May/20 03:03

Description

The use of read_feather function with CompressedInputStream fails with the error:

 

Error in ipc___feather___Reader__Open(file) : 
 Invalid: File is too small to be a well-formed file

The minimal code to get it:

filename <- "x.gz"
data <- data.frame(c(1,2), c(3,4), stringsAsFactors=FALSE)
fos <- FileOutputStream$create(filename)
codec <- Codec$create("gzip")
cos <- CompressedOutputStream$create(fos, codec)
write_feather(data, cos)
cos$close()
fos <- ReadableFile$create(filename)
codec <- Codec$create("gzip")
cos <- CompressedInputStream$create(fos, codec)
read_feather(cos)

However, I can extract the file "x.gz" with

gzip -d x.gz

and then run in R:

fos <- ReadableFile$create("x")
read_feather(fos);
  c.1..2. c.3..4.
1       1       3
2       2       4

The codec gzip is available according to:

codec_is_available("gzip")
[1] TRUE

Attached the generated "x.gz"

 

Issue Links

• links to: GitHub Pull Request #7185