Python re to extract useful information from each line

Tim Chase python.list at tim.thechases.com
Thu Apr 30 00:38:57 CEST 2015


On 2015-04-29 14:22, Emile van Sebille wrote:
> On 4/29/2015 1:49 PM, Kashif Rana wrote:
> > pol_elements =
> > re.compile('id\s(?P<p_id>.+?)(?:\sname\s(?P<p_name>.+?))?\sfrom\s(?P<p_from>.+?)\sto\s(?P<p_to>.+?)\s{2}(?P<p_src>[^\s]+?)\s(?P<p_dst>[^\s]+?)\s(?P<p_port>[^\s]+?)(?:(?P<p_nat_status>\snat)\s(?P<p_nat_type>[^\s]+?)(?P<p_nat_ip>\sdip-id\s[^\s]+?)?)?\s(?P<p_action>[^\s]+?)(?:\sschedule\s(?P<p_schedule>[^\s]+?))?(?P<p_log_status>\slog)?$'
> > )
> 
> ... and that's why we avoid regular expressions... it makes my head
> hurt just looking at that line noise.

First, it appears the OP isn't using raw strings which make those
back-slashes just ask for trouble.

That said, it would be a lot better if the OP made use of re.VERBOSE
to put each component on its own line:

 pol_elements = re.compile(r"""
   id
   \s
  (?P<p_id>.+?)
  (?:
    \s
    name
    \s
    (?P<p_name>.+?)
    )?
  \s
  from
  \s
  (?P<p_from>.+?)
  \s
  to
  \s
  (?P<p_to>.+?)
  \s{2}
  (?P<p_src>[^\s]+?)
  \s
  (?P<p_dst>[^\s]+?)
  \s(?P<p_port>[^\s]+?)
  (?:
    \s
    (?P<p_nat_status>nat)
    \s
    (?P<p_nat_type>\w+)
    (
      \s?
      P<p_nat_src_ip>dip-id
      \s
      \d+   
      )?
    (
      \s
      ip
      \s
      (?P<p_nat_dst_ip>[\d\.]+)
      \s
      port
      (?P<dst_nat_port>\d+)
      )?
    )?
  \s
  (?P<p_action>[^\s]+?)
  (?:
    \s
    schedule
    \s
    (?P<p_schedule>[^\s]+?)
    )?
  (?P<p_log_status>\slog)?
  $
   """, re.VERBOSE)

which, with some copious comments in the expression, would make it
almost readable.

Alternatively, switch to an actual parser like pyparsing.

-tkc






More information about the Python-list mailing list