Split String

[Kemal]

I have a strings that includes mail address with ';'
separetor so I need to split this string into substrings or
into an array like in
C#
Type string[];
Type = str.Split(';');
How can i do this in powerbuilder? Is there a function like
that or i need to write it myself?

[Terry Dykstra]

Here is some code I found a long time ago.

global function integer gf_parse_plus (string as_string, ref string
as_array[], ref integer ai_count, string as_delimiter);// Function Takes a
string and separates elements into an array

// Returns 0 = success
// 1 = error
//
// as_string Source string containing values delimited by as_delimiter
// as_array[] Destination array, passed by value
// ai_count Returns count of items in array
// as_delimiter delimiter value

int li_pos // position of commas
int li_len // string parsing
int li_start // string parsing
string ls_item // single item parsed from string

if Len ( as_string ) < 1 then
ai_count = 0
return 1
end if

// parse out items and insert into array

li_pos = 1
li_start = 1
ai_count = 0

DO WHILE li_pos > 0

li_pos = Pos ( as_string, as_delimiter, li_start )

if li_pos = 0 then
ls_item = Mid ( as_string, li_start )
else
ls_item = Mid ( as_string, li_start, li_pos - li_start )
end if

ls_item = Trim ( ls_item )

if Len ( ls_item ) > 0 then
ai_count++
as_array[ai_count] = ls_item
end if

li_start = li_pos + 1

LOOP

return 0
--
Terry Dykstra (TeamSybase)
http://powerbuilder.codeXchange.sybase.com/
http://casexpress.sybase.com/cx/cx.stm
product enhancement requests:
http://my.isug.com/cgi-bin/1/c/submit_enhancement

<Kemal> wrote in message news:4803885f.644...@sybase.com...

[Jeremy Lakeman]

I find most of the time you don't really need an array since you intend
to process the data from start to finish anyway. I often found myself
writing code like the following to parse strings;

do while str<>''
value=left(str,pos(str+';',';')-1)
str=mid(str,pos(str+';',';')+1)

// process value....

loop

So at some point I generalised that coding pattern and wrote a helper
function;

global function boolean f_token (ref string as_source, string as_delim,
ref string as_value);
boolean lb_ret
long ll_pos

lb_ret=as_source<>''
ll_pos=Pos (as_source + as_delim, as_delim)
as_value = trim (Left (as_source, ll_pos - 1))
as_source = trim (Mid (as_source, ll_pos + Len (as_delim)))
return lb_ret
end function

Then I overloaded it with different data types to convert the value
automatically;

global function boolean f_token (ref string as_source, string as_delim,
ref decimal adc_value)
global function boolean f_token (ref string as_source, string as_delim,
ref datetime adtm_value)
global function boolean f_token (ref string as_source, string as_delim,
ref boolean ab_value)
...

(conversion and error handling left as an exercise for the reader)

So now I end up with code that looks like this;

ls_file = filereadex( ....

do while f_token(ls_file, '~r~n', ls_line)

f_token(ls_line, ',', ldc_key)
f_token(ls_line, ',', ls_name)

...

insert ...
values (:ldc_key, :ls_name, ...
loop