| Subject: |
RE: [TME10][TEC] Tec predicate similar to perl split |
I messed up the cut and paste:
The re_matches in the first example should be:
re_match(re_1,_msg,1,_re_date),
re_match(re_1,_msg,2,_re_time),
re_match(re_1,_msg,3,_re_host),
re_match(re_1,_msg,4,_re_path),
re_match(re_1,_msg,5,_re_app),
re_match(re_1,_msg,6,_re_data)
-----Original Message-----
From: Blankenship, I.V. (Contractor)
[mailto:[email protected]]
Sent: Thursday, July 29, 2004 9:08 AM
To: '[email protected]'
Subject: RE: [tme10] [TME10][TEC] Tec predicate similar to perl split
I am surprised that old TEC 3.6 stuff is still floating around ;-)
I think that the re_* regular expression predicates provided since 3.7 will
work better and with less overhead.
For the specific problem in the original post this should work:
reception_action: tec_start_action:
(
% Date time host path
app data
re_create(re_1,'^\s*([^\|]+)\s*\|\s*([^\|]+)\s*\|\s*([^\|]+)\s*\|\s*([^\|]+)
\s*\|\s*([^\|]+)\s*\|\s*(.+)$'),
)
In the processing rule I assume that the msg slot will be split:
action: split_msg:
(
re_match(re_1,_msg,1,_re_date),
re_match(re_2,_msg,1,_re_time),
re_match(re_3,_msg,1,_re_host),
re_match(re_4,_msg,1,_re_path),
re_match(re_5,_msg,1,_re_app),
re_match(re_6,_msg,1,_re_data)
)
If I wanted to make a more generic predicate that worked along the lines of
the perl split I could do something such as this:
Step 1 - define the Prolog rules to split the string:
split(_regex_id,_string,_list) :-
atom(regex_id),atom(_string),
re_split(_regex_id,_string,_list),!.
re_split(_regex_id,_string,_list) :-
re_before_match(_regex_id,_string,_element),
(re_after_match(_regex,_id_string,_rest) ->
(re_split(_regex_id,_rest,_working),!);_working=[]),
_list=[_element|_working].
re_split(_,_string,[_string]).
Step 2 - setup regular expression and processing rule:
reception_action: tec_start_action:
(
re_create(re_1,'\s*\|\s*'),
)
action: split_msg:
(
split(re_1,_msg,[_re_date,_re_time,re_path,_re_app,_re_data])
)
I have not tested the code listed above so it may need some tweaking.
-----Original Message-----
From: [email protected]
[mailto:[email protected]]
Sent: Thursday, July 29, 2004 7:05 AM
To: [email protected]
Subject: Re: [tme10] [TME10][TEC] Tec predicate similar to perl split
I have not been around for the last couple of days so this may have been
answered already. Anyway, here is I.V.'s original predicate for splitting a
string. At the moment, I cannot find his upgraded, more efficient version.
reception_action: split_atom:
(
assert((
split_atom(ATOM,DELIM,_LIST) :-
nonvar(ATOM),nonvar(DELIM),var(LIST),
name(ATOM,_codes),
name(DELIM,_list),
[_delim|[]]=_list,
re_split_atom(_codes,_delim,[],_working),!,
_LIST=_working
)),
assert((
re_split_atom([],_,[],_WORKING):- _WORKING=[]
)),
assert((
re_split_atom([],_,CURRENT,_WORKING) :-
name(_current,CURRENT),
_WORKING=[_current]
)),
assert((
re_split_atom([DELIM|Y],DELIM,[],_WORKING) :-
re_split_atom(Y,DELIM,[],_WORKING),!
)),
assert((
re_split_atom([DELIM|Y],DELIM,CURRENT,_WORKING) :-
re_split_atom(Y,DELIM,[],_working),!,
name(_current,CURRENT),
_WORKING=[_current|_working]
)),
assert((
re_split_atom([X|Y],DELIM,[],_WORKING) :-
re_split_atom(Y,DELIM,[X],_WORKING),!
)),
assert((
re_split_atom([X|Y],DELIM,CURRENT,_WORKING) :-
append(CURRENT,[X],_current),
re_split_atom(Y,DELIM,_current,_working),!,
_WORKING=_working
))
),
Kyle Amundson
3M Tivoli Group
224-3W-17
Phone: 651-575-1906
Fax: 651-736-3125
john_willis@gulfso
ft.com To:
[email protected]
Sent by: cc:
owner-tme10@lists. Subject: Re: [tme10]
[TME10][TEC] Tec predicate similar to perl split
us.ibm.com
07/29/2004 05:20
AM
Please respond to
tme10
Look at the re_create and re_match predicates.
Jim Sander did an article on RegEx predicates in this newsletter:
http://www.gulfsoft.com/downloads/GBF_Newsv1_3.pdf
John Willis
Gulf Breeze Software
www.gulfsoft.com
Quoting "Haynes Ian (UK)" <[email protected]>:
> List
>
>
>
> Does anyone know of a tec predicate similar in functionality to the perl
> split function.
>
>
>
> I have slot which is pipe "|" delimited and would like to split it up int
to
> its 6 constituent parts
>
>
>
> Ie
>
>
>
> Date|time|host|path|application|data
>
>
>
> The information is not always the same length so I need to be able to
split
> the slot at the | symbol and put the data into a variable
>
>
>
>
>
> Many thanks
>
>
>
> Ian Haynes
Use the -f option to specify a config file. This file should be an EIF compliant configuration file so you can use the BufEvtPath=pathname
keyword to specify a cache file.
