Bug #8784: CSV - Empty fields are discarded when col_sep is a space

• Status: Closed
• Assignee: kou (Kouhei Sutou)
• Target version: -
• ruby -v: ruby 2.0.0p247 (2013-06-27 revision 41674) [x86_64-linux]

[grim7reaper (Sylvain Laperche)]

When using space as column delimiter, empty fields are discarded.
With other delimiters, like comma, empty fields are correctly retrieved.

The following code reproduces the problem:

---

#!/usr/bin/env ruby

require 'csv'

print(CSV.parse('2009,2,3,8,0,,30.1,,'))
puts
print(CSV.parse('2009 2 3 8 0 30.1 ', col_sep: ' '))
puts

---

Expected output
[["2009", "2", "3", "8", "0", nil, "30.1", nil, nil]]
[["2009", "2", "3", "8", "0", nil, "30.1", nil, nil]]

Current output
[["2009", "2", "3", "8", "0", nil, "30.1", nil, nil]]
[["2009", "2", "3", "8", "0", "30.1", nil]]

[grim7reaper (Sylvain Laperche)]

• File bug_csv_col_sep.rb bug_csv_col_sep.rb added

The bug is still present in Ruby 2.1.0

% ruby -v
ruby 2.1.0p0 (2013-12-25 revision 44422) [x86_64-linux]

The attached script prints:

[["2009", "2", "3", "8", "0", nil, "30.1", nil, nil]]
[["2009", "2", "3", "8", "0", "30.1", nil]]

instead of:

[["2009", "2", "3", "8", "0", nil, "30.1", nil, nil]]
[["2009", "2", "3", "8", "0", nil, "30.1", nil, nil]]

[hsbt (Hiroshi SHIBATA)]

• Status changed from Open to Assigned
• Assignee set to JEG2 (James Gray)

[slavcho42 (Slavcho Ivanov)]

The problem is in the String::split function.
It turns out it is a feature, not a bug: http://ruby-doc.org/core-2.4.0/String.html#method-i-split .
"If pattern is a single space, str is split on whitespace, with leading whitespace and runs of contiguous whitespace characters ignored."

This should fix the problem in csv.rb - line 1834:

old:
parts = parse.split(@col_sep, -1)

new:
sep = @col_sep == ' ' ? / / : @col_sep
parts = parse.split(sep, -1)

Hope this helps.

[mame (Yusuke Endoh)]

• Assignee changed from JEG2 (James Gray) to kou (Kouhei Sutou)
• Backport deleted (1.9.3: UNKNOWN, 2.0.0: UNKNOWN)

@kou (Kouhei Sutou), could you check this ticket?

[kou (Kouhei Sutou)]

• Status changed from Assigned to Closed

Thanks for your report.
I've fixed it at the master: https://github.com/ruby/csv/commit/ba560e407a152afffea589d832084c249471eeb6