Mirror of forum.nim-lang.org

5444 :: Regex extracting string between two words not working

• mronetwo (orginal) [2019-11-03T12:34:28+01:00] view original

  Hello,

  I hope it won't be a garbage post, because I'm missing something super simple...

  I'm sending a Json from a Python script to a server written in Nim. The Json data is between START and END strings.

  if "START" in buf:
    buildingRecord = true
  
  if buildingRecord:
    dataBuf.add(buf)
  
  if "END" in buf and buildingRecord:
    # Extract the message.
    var matches: seq[string]
    var matched = find(dataBuf, re"START((.|\s)*)END", matches)
    
    if matched > 0:
      echo(matches[0])
  Run

  The regex finds no matches. I hope it's not a problem with the regex itself. I've tested it online and it seems to work. The Json looks like so:

  {
    "start_h":10,
    "start_m":40,
    "end_h":11,
    "end_m":40,
    "delta_h":0,
    "delta_m":0,
    "project_ID":0
  }
  Run

  It gets wrapped between START and END and pushed through a socket to the server. Server gets the data for sure. It just fails to extract the data with regex.

---

• libraM (orginal) [2019-11-03T19:21:00+01:00] view original

  With some fixes:

  import re, strutils
  
  let buf = """
  START
  {
    "start_h":10,
    "start_m":40,
    "end_h":11,
    "end_m":40,
    "delta_h":0,
    "delta_m":0,
    "project_ID":0
  }
  END
  """
  
  var
    dataBuf: string = ""
    buildingRecord: bool
  
  if "START" in buf:
    buildingRecord = true
  
  if buildingRecord:
    dataBuf.add(buf)
  
  if "END" in buf and buildingRecord:
    # Extract the message.
    var matches: array[1, string]
    var matched = find(dataBuf, re(r"START(.+)END", flags={reDotAll}), matches)
    
    if matched >= 0:
    # Or just:
    # if contains(dataBuf, re(r"START(.+)END", flags={reDotAll}), matches):
      echo(matches[0])
  Run

---

• mronetwo (orginal) [2019-11-04T18:33:21+01:00] view original

  Thank you. That works. Could you explain to me what was wrong with my code? It seems that the important part is the

  var matches: array[1, string]
  Run

  Why is

  var matches: seq[string]
  Run

  wrong? I thought that's a list of strings that can be dynamically extended.

---

• pietroppeter (orginal) [2020-12-03T15:10:54+01:00] view original

  bumping this in case you are still wondering after 1 year...

  the answer is that using seq does not work in re.match and you have to know beforehand the exact length of the matches array. This is an acknowledge bug (which nre does not have): https://github.com/nim-lang/Nim/issues/9472

  (see also this SO answer)

---