How to parse a part of a link in python?

[pmasiar]

Compare: regex solution:

/[?&]v=([^&]+)/.match url
result = $1 

vs plain string parsing:

result = url.split('v=')[1] 
if '&' in result:
    result = result.split('&')[0] 

It takes one more line, but is easier to read, unless you are regex expert.

Of course regex solution scales for more complex cases. But in 80% cases, like this one, regex is overkill, IMHO. YMMV, eat your regex if you want, I don't care anymore.

[slavik]

Of course regex solution scales for more complex cases. But in 80% cases, like this one, regex is overkill, IMHO. YMMV, eat your regex if you want, I don't care anymore.

Once I get a regex to match correctly, I don't have to care about it anymore.

[ghostdog74]

Once I get a regex to match correctly, I don't have to care about it anymore.

that's quite untrue isn't it? how do you know your program specs won't change in future?

[WW]

The structure is simple enough; regex does seem like overkill. This python function gets all the assigned values into a dict. Python gurus could probably simplify it even more:

# parse_url_opts.py

def parse_url_opts(str):
    (url,opts_str) = str.split("?")
    opts_list = opts_str.split("&")
    opts_pairs = [s.split("=") for s in opts_list]
    return dict(opts_pairs)

Example:

>>> from parse_url_opts import *
>>> z = parse_url_opts("http://www.youtube.com/watch?v=ngjr6IkDLxg&feature=related")
>>> z.keys()
['feature', 'v']
>>> z['feature']
'related'
>>> z['v']
'ngjr6IkDLxg'
>>>

(Something like this might already exist in the cgi module.)

[mssever]

if a simple regex handles all variations well, it most probably would not be simple anymore, don't you agree?

Not necessarily. The regex I wrote above is quite simple.

[ghostdog74]

Not necessarily. The regex I wrote above is quite simple.

yes for this simple case. like you mentioned, if the url contains different kinds of query strings and not just standard "v=xxxxxx", then a simple regex can become quite complex if all variations are to be handled.

[slavik]

# parse_url_opts.py

def parse_url_opts(str):
    (url,opts_str) = str.split("?")
    opts_list = opts_str.split("&")
    opts_pairs = [s.split("=") for s in opts_list]
    return dict(opts_pairs)

the Perl version:

sub parse_url_opts {
  my $url = shift;
  (my $base, my $opts) = split /\?/, $url;
  my @opt_list = split /&/, $opts;
  my %opt_pairs = map { split /=/, $_ } @opt_list;
  return %opt_pairs;
}

the only real difference is how the subroutines get their arguments and in Perl, the map is explicit.

[LaRoza]

the Perl version:

sub parse_url_opts {
  my $url = shift;
  (my $base, my $opts) = split /\?/, $url;
  my @opt_list = split /&/, $opts;
  my %opt_pairs = map { split /=/, $_ } @opt_list;
  return %opt_pairs;
}

the only real difference is how the subroutines get their arguments and in Perl, the map is explicit.

There is also a "map" function in Python. Although I don't normally use it, it may be used in a way like yours (I just got up and didn't really read the code, except to say the Perl example is very different visually from the Python)

[pmasiar]

map will be dropped in Py3K IIRC. List comprehension does the same in a more obvious and readable way.

[LaRoza]

map will be dropped in Py3K IIRC. List comprehension does the same in a more obvious and readable way.

http://www.artima.com/weblogs/viewpost.jsp?thread=98196

Update: lambda, filter and map will stay (the latter two with small changes, returning iterators instead of lists). Only reduce will be removed from the 3.0 standard library. You can import it from functools.