What I need to do is convert something like this:
"id,name,user[id,email]"
into this array:
["id", "name", {"user"=>["id", "email"]}]
What is the best way to do that? I think some function like split or scan can help, but I don't have much knowledge in regex to solve this.
Just out of curiosity:
▶ str = "id,name,user[id,email]"
▶ eval "[#{str.gsub(/(\w+)\[(.*?)\]/, '{\1=>[\2]}').gsub(/\w+/, ':\0')}]"
#⇒ [
# [0] :id,
# [1] :name,
# [2] {
# :user => [
# [0] :id,
# [1] :email
# ]
# }
#]
Disclamer: to use eval in production one must understand risks.
UPD Safe evaling (note that every ASCII \w symbol is converted to it’s wide pair from UTF-8 to prevent injection; not the best way around, but it works nicely unless you have ruby functions named with wide characters):
▶ safe = str.gsub(/\w/) do |e|
▷ e.each_codepoint.map do |cp|
▷ cp + 0xFF00 - 0x0020
▷ end.pack('U')
▷ end
#⇒ "ｉｄ,ｎａｍｅ,ｕｓｅｒ[ｉｄ,ｅｍａｉｌ]"
▶ eval "[#{safe.gsub(/(\p{L}+)\[(.*?)\]/, '{\1=>[\2]}').gsub(/\p{L}+/, ':\0')}]"
#⇒ [
# [0] :ｉｄ,
# [1] :ｎａｍｅ,
# [2] {
# :ｕｓｅｒ => [
# [0] :ｉｄ,
# [1] :ｅｍａｉｌ
# ]
# }
#]
Now you are free to turn keys back from wide characters to ASCII.
safe I'm getting: undefined local variable or method ｎａｍｅ'`irb it does not happen. Will figure it out. Thanks.@ may appear in email etc, it should be considered as a proof of concept only.\w would not work with wides neither in rails nor in plain ruby. There is \p{L} matcher for utf-8 word characters. ruby-doc.org/core-2.2.1/…string = "id,name,user[id,email]"
parts = string.split(/[\,,\[,\]]/)
[parts[0], parts[1], {parts[2] => [parts[3], parts[4]]}]
You may also want to have some validation or not use all the parts if the first id is supposed to be the same as the second.
"user[id,email],id,name"? The order can change, the number of strings can change, everything can change, my example is just a pattern. The @mudasobwa answer solves my problem but I did not want to use eval.
eval is not an evil. It might be used when you understand the circumstances. Absolutely safe usage: convert all \w to some utf-8, like wide characters ⇒ do eval ⇒ convert everything back.
["id", "name", "user" => ["id", "email"]]to["id", "name", {"user"=>["id", "email"]}]. In the example, user became a hash inside the array. You can try it typing my original array in console.