How to write a regex in a single line - ruby

Question

I have this code:
str = 'printf("My name is %s and age is %0.2d", name, age);'
SPECIFIERS = 'diuXxofeEgsc'
format_specifiers = /((?:%(?:\*?([-+]?\d*\.?\d+)*(?:[#{SPECIFIERS}]))))/i
variables = /([.[^"]]*)\);$/
format = str.scan(format_specifiers)
var = str.scan(variables).first.first.split(/,/)
Is there any way a single regex can do that in a couple of lines?
My desired output is:
%s, name
%0.2d, age

Answer 1

I'm a big believer in keeping regular expressions as simple as possible; They can too quickly mushroom into unwieldy/unmaintainable messes. I'd start with something like this, then tweak as necessary:
str = 'printf("My name is %s and age is %0.2d", name, age);'
formats = str.scan(/%[a-z0-9.]+/) # => ["%s", "%0.2d"]
str[/,(.+)\);$/] # => ", name, age);"
vars = str[/,(.+)\);$/].scan(/[a-z]+/) # => ["name", "age"]
puts formats.zip(vars).map{ |a| a.join(', ')}
# >> %s, name
# >> %0.2d, age

Answer 2

Your question has two parts:
Q1: Is it possible to do this with a single regex?
Q2: Can this be done in one or two lines of code?
The answer to both questions is "yes".
format_specifiers = /
%[^\s\"\z]+ # match % followed by > 0 characters other than a
# whitespace, a double-quote or the end of the string
/x # free-spacing regex definition mode
variables = /
,\s* # match comma followed by >= 0 whitespaces
\K # forget matches so far
[a-z] # match a lowercase letter
\w* # match >= 0 word characters
/x
You can decide, after testing, if these two regexes do their jobs adequately. For testing, refer to Kernel#sprintf.
r = /
(?:#{format_specifiers}) # match format_specifiers in a non-capture group
| # or
(?:#{variables}) # match variables in a non-capture group
/x
#=> /
(?:(?x-mi:
%[^\s\"\z]+ # match % followed by > 0 characters other than a
# whitespace, a double-quote or the end of the string
)) # match format_specifiers in a non-capture group
| # or
(?:(?x-mi:
,\s* # match comma followed by >= 0 whitespaces
\K # forget matches so far
[a-zA-Z] # match a letter
\w* # match >= 0 word characters
)) # match variables in a non-capture group
/x
r can of course also be written:
/(?:(?x-mi:%[^\s\"\z]+))|(?:(?x-mi:,\s*\K[a-zA-Z]\w*))/
One advantage of constructing r from two regexes is that each of the latter can be tested separately.
str = 'printf("My name is %s and age is %0.2d", name, age);'
arr = str.scan(r)
#=> ["%s", "%0.2d", "name", "age"]
arr.each_slice(arr.size/2).to_a.transpose.map { |s| s.join(', ') }
#=> ["%s, name", "%0.2d, age"]
I have five lines of code. We could reduce this to two by simply substituting out r in str.scan(r). We could make it a single line by writing:
str.scan(r).tap { |a|
a.replace(a.each_slice(a.size/2).to_a.transpose.map { |s| s.join(', ') }) }
#=> ["%s, name", "%0.2d, age"]
with r substituted out.
The steps here are as follows:
a = str.scan(r)
#=> ["%s", "%0.2d", "name", "age"]
b = a.each_slice(a.size/2)
#=> a.each_slice(2)
#=> #<Enumerator: ["%s", "%0.2d", "name", "age"]:each_slice(2)>
c = b.to_a
#=> [["%s", "%0.2d"], ["name", "age"]]
d = c.transpose
#=> [["%s", "name"], ["%0.2d", "age"]]
e = d.map { |s| s.join(', ') }
#=> ["%s, name", "%0.2d, age"]
a.replace(e)
#=> ["%s, name", "%0.2d, age"]
The methods used (aside from Array#size) are String#scan, Enumerable#each_slice, Enumerable#to_a, Enumerable#map, Array#transpose and Array#replace.