How to extract the last value before last : character but after the last comma?

I'm using Clojure to parse and analyze XML files.

Here is a sample:

BSS:17,NSVC:1
BSS:17,NSVC:4
BSS开发者_如何转开发:17,NSVC:5
BSS:17,BTSM:0,BTS:3
BSS:17,BTSM:0,BTS:4
BSS:17,BTSM:0,BTS:5
BSS:17,BTSM:1,BTS:0
BSS:17,BTSM:1,BTS:1
BSS:17,BTSM:1,BTS:2
BSS:17,BTSM:1,BTS:3

I'm interested in that last value (a value after the last comma but before the last : , NSVS and BTS in my case), digits after them don't matter.

How to extract that last value in the previous strings?

---

You can use this function to process the individual lines:

(defn lastval [s]
  (second (re-find #",([^,:]+):\d*$" s)))
                 ;   ^ the comma preceding the interesting section
                 ;    ^ the part in parens will be captured as a group
                 ;     ^ character class meaning "anything except , or :"
                 ;            ^ the colon after the interesting section
                 ;             ^ any number of digits after the colon
                 ;                ^ end of string
          ; ^ returns a vector of [part-that-matches, first-group];
          ;   we're interested in the latter, hence second

NB. this returns nil if the regex doesn't match.

E.g.:

user> (lastval "BSS:17,BTSM:0,BTS:3")
"BTS"

If you later want to extract all the information in easy-to-work-with bits, you can use

(defn parse [s]
  (map (juxt second #(nth % 2)) (re-seq #"(?:^|,)([^,:]+):(\d+)" s)))

E.g.

user> (parse "BSS:17,BTS:0,BTS:3")
(["BSS" "17"] ["BTS" "0"] ["BTS" "3"])

---

Can you use lastIndexOf to find the last comma?

---

Does this work for you?

(def tmp (str "BSS:17,NSVC:1\n
BSS:17,NSVC:4\n
BSS:17,NSVC:5\n
BSS:17,BTSM:0,BTS:3\n
BSS:17,BTSM:0,BTS:4\n
BSS:17,BTSM:0,BTS:5\n
BSS:17,BTSM:1,BTS:0\n
BSS:17,BTSM:1,BTS:1\n
BSS:17,BTSM:1,BTS:2\n
BSS:17,BTSM:1,BTS:3\n"))


(defn split [s sep]
  (->> (.split s sep)
       seq
       (filter #(not (empty? %)))))

(reduce (fn[h v]
          (conj h (last v)))

          [] (map #(split % ",")
                  (split tmp "\n")))

I am assuming there is some sort of delimeter between the lines.