Question

I need a regular expression that matches this text:

894975||0||Lever 2000 Anti-Bacterial Bar Soap ||Health & Spa/Personal Care/Body Care/Soap

I want to search the text, and, after two sets of pipes, match “Bar Soap”.

If the words are not in order then it’s not matching.
My regex is:

/^(?:\d+\|\|).?\|\|[^|]*?(Bar[^|]*? Soap)/i 

This is not matching when “soap” comes first and “bar” second.

Answer 1

The sample data looks like a standard pipe-delimited ('|') file you’d see from database extracts. It’s common to see fields with a null value show up as || in the output.

Rather than try to parse it using a regex, it’s usually handled by splitting on the pipes, or by treating it as a CSV record with a pipe instead of comma as the column-separator. Splitting on double-pipes (||) will fail if you get a record where the field actually has content.

Here’s two different samples showing how I’d do it. The first is splitting on | into fields.

text = '894975||0||Lever 2000 Anti-Bacterial Bar Soap ||Health & Spa/Personal Care/Body Care/Soap'
fields = text.split('|')

fields at this point looks like:

[
    [0] "894975",
    [1] "",
    [2] "0",
    [3] "",
    [4] "Lever 2000 Anti-Bacterial Bar Soap ",
    [5] "",
    [6] "Health & Spa/Personal Care/Body Care/Soap"
]

Grabbing the fifth field retrieves the product:

product = fields[4]
=> "Lever 2000 Anti-Bacterial Bar Soap" 

A second way to do this is to treat the content as a CSV file with | delimiters:

require 'csv'

text = <<EOT
894975||0||Lever 2000 Anti-Bacterial Bar Soap ||Health & Spa/Personal Care/Body Care/Soap
EOT

CSV.parse(text, :col_sep => '|') do |row|
  puts row[4]
end

=> "Lever 2000 Anti-Bacterial Bar Soap" 

The advantage of doing it using CSV is it’s possible to find a | character in text, and CSV will handle decoding embedded pipes correctly.

Because there’s only one sample input line, this solution can’t be more thorough.