CODEWITHSUNDEEP
regexerlang
Alec Smart
Alec Smart

Reputation: 95880

Remove Whitespace using Erlang Regex

I want to remove all the whitespace i..e tabs/spaces/newline chars.

T = {xmlelement,"presence",
                                         [{"xml:lang","en"}],
                                         [{xmlcdata,<<"\n">>},
                                          {xmlelement,"priority",[],
                                              [{xmlcdata,<<"5">>}]},
                                          {xmlcdata,<<"\n">>},
                                          {xmlelement,"c",
                                              [{"xmlns",
                                                "http://jabber.org/protocol/caps"},
                                               {"node","http://psi-im.org/caps"},
                                               {"ver","0.12.1"},
                                               {"ext","cs ep-notify html"}],
                                              []},
                                          {xmlcdata,<<"\n">>}]}.

I tried the following, but it does not work:

trim_whitespace(Input) ->
re:replace(Input, "(\r\n)*", "").

Upvotes: 1

Views: 5863

Answers (4)

xgMz
xgMz

Reputation: 3344

re:replace is tricky, something to keep in mind:

Eshell V5.9.3.1  (abort with ^G)
1> re:replace("0 1  2    3 4 5 6 7 8 9", " ", "", [global, {return, list}]).
"0123456789"
2> re:replace("0 1  2    3 4 5 6 7 8 9", " ", "", [{return, list}]).
"01  2    3 4 5 6 7 8 9"
3> re:replace("0 1  2    3 4 5 6 7 8 9", " ", "").
[<<"0">>,[]|<<"1  2    3 4 5 6 7 8 9">>]

Upvotes: 0

fenollp
fenollp

Reputation: 2496

I faced the same issue… came here to share my more efficient work:

trim(Subject) ->
  {match, [[Trimmed]|_]} = re:run(Subject, "^\\s*([^\\s]*(?:.*[^\\s]+)?)\\s*$", 
    [{capture, all_but_first, binary}, global, dollar_endonly, unicode, dotall]),
  Trimmed.

The idea is very much the same. The regex is just better.

Upvotes: 2

archaelus
archaelus

Reputation: 7129

All the whitespace in your question is in cdata sections - why not just filter those out of the tuple?

remove_cdata(List) when is_list(List) ->
    remove_list_cdata(List);
remove_cdata({xmlelement, Name, Attrs, Els}) ->
    {xmlelement, Name, remove_cdata(Attrs), remove_cdata(Els)}.

remove_list_cdata([]) ->
    [];
remove_list_cdata([{xmlcdata,_}|Rest]) ->
    remove_list_cdata(Rest);
remove_list_cdata([E = {xmlelement,_,_,_}|Rest]) ->
    [remove_cdata(E) | remove_list_cdata(Rest)];
remove_list_cdata([Item | Rest]) ->
    [Item | remove_list_cdata(Rest)].


remove_cdata(T) =:= 
    {xmlelement,"presence",
     [{"xml:lang","en"}],
     [{xmlelement,"priority",[],[]},
      {xmlelement,"c",
       [{"xmlns","http://jabber.org/protocol/caps"},
        {"node","http://psi-im.org/caps"},
        {"ver","0.12.1"},
        {"ext","cs ep-notify html"}],
       []}]}

Upvotes: 0

Matt K
Matt K

Reputation: 13842

If you want to remove everything in a string, you need to pass the global option to re:replace(). You're also only replacing newlines by using that regex. The call should probably look like this:

trim_whitespace(Input) -> re:replace(Input, "\\s+", "", [global]).

Upvotes: 4

Related Questions