Extracting from String

[former_member5350]

Hi,

if i have a String like 'M[ey]er' how could i extract all information that is not within the brackets so it would give me 'Mer' then ?

Thanks,

[former_member5350]

Hi,

Following is the code in addition to what all has been said by others

data : string(20),

string1(20),

string2(20),

string3(20),

string4(20),

del value '[',

del2 value ']'.

string = 'M[ey]er'.

split string at del into string1 string2.

*write / string1.

split string2 at del2 into string3 string4.

*write / string4.

concatenate string1 string4 into string1.

write / string1. "Gives you Mer

[Former Member]

u can use keyword

SPLIT 'PRABHU' at 'R' into x1 x2 .

http://help.sap.com/saphelp_47x200/helpdata/en/fc/eb3400358411d1829f0000e829fbfe/frameset.htm

after that u can use concatenate keyword to get into single field.

Regards

Prabhu

[Former Member]

SPLIT string at '[' into x1 x2.

SPLIT x2 at ']' into x3 x4.

concatenate x1 x4 into x5 then x5 will contain the required string 'Mer'

[former_member5350]

thank you, here is what i got:

 DATA:
   w_string TYPE string ,
   w_sub_string TYPE string,
   w_int TYPE i,
   zaehl TYPE i,
   w_char,
   w_counter TYPE i,
   fl_flag.

  w_string = i_s2.

  w_int = STRLEN( w_string ).
  w_counter = 0.

  DO w_int TIMES.
    zaehl = zaehl + 1.
    w_char = w_string+w_counter(1).

    IF w_char EQ '['.
      zaehl = zaehl - 1.
      WRITE:/ zaehl.
  
      CLEAR fl_flag.
    ENDIF.


    IF fl_flag = 'X'.
      CONCATENATE  w_sub_string w_char INTO w_sub_string.
    ENDIF.

    IF w_char EQ ']'.
      zaehl = 0.
      fl_flag = 'X'.

    ENDIF.

    w_counter = w_counter  + 1.
   
  ENDDO.
  
 WRITE:/ zaehl.
  i_s2 = w_sub_string.

ENDFORM.

Now how do i have to adapt the code that i can do the following:

If i have the string 'Jose[ph]M[ue]ller' then i want to get '[][ph][?][ue][]'.

I will surely reward points generously!

[Former Member]

data str type string.

data str1 type string.

data str2 type string.

data str3 type string.

data str4 type string.

data str5 type string.

str = 'p[riya]nk'.

split str at '[' into str1 str2.

write : / str1, str2.

split str2 at ']' into str3 str4.

write 😕 str3, str4.

if strlen( str1 ) = 1.

str1 = '?'.

else.

str1 = '*'.

endif.

if strlen( str4 ) = 1.

str4 = '?'.

else.

str4 = '*'.

endif.

concatenate str1 '[' str3 ']' str4 into str5.

write 😕 str5.

Message was edited by:

Priyank Jain

[Former Member]

HI..

Just paste this code..

data:

w_string type string value 'ram[mohan]hh[rao]kk[nagam]',

w_len type i,

w_cnt type i.

data:

w_output type string,

w_temp type string.

data:

itab like standard table of w_string with header line.

replace all occurrences of '[' in w_string with ']'.

<b>split w_string at ']' into table itab.</b>

<b>loop at itab .</b>

w_cnt = sy-tabix mod 2.

if w_cnt eq 1.

w_temp = '[*]'.

else.

w_temp = itab.

endif.

<b>concatenate w_output w_temp into w_output.</b>

clear w_temp.

<b>endloop.</b>

<b>write: w_output.</b>

[former_member5350]

Thank you, but i am on 4.6 so "all occurrences" wont work, sorry!

[Former Member]

Hi..

try with this..<b>without replace</b> stmt..

data:

w_string(50) type C value 'ram[mohan]hh[rao]kk[nagam]',

w_len type i,

w_cnt type i.

data:

w_output type string ,

w_temp type string.

data:

itab like standard table of w_string with header line.

w_len = strlen( w_string ).

do w_len times.

w_cnt = sy-index - 1.

if w_string+w_cnt(1) eq '['.

w_string+w_cnt(1) = ']'.

endif.

enddo.

<b>split w_string at ']' into table itab.</b>

<b>loop at itab .</b>

w_cnt = sy-tabix mod 2.

if w_cnt eq 1.

w_temp = '[*]'.

else.

w_temp = itab.

endif.

concatenate w_output w_temp into w_output.

clear w_temp.

<b>endloop.

write w_output.</b>

[former_member5350]

Hi Rammohan,

this is quite close to what i need, thank you! Could you adjust it once more for me ?

What i need is when i have 'ram[mohan]hh[rao]k[nagam]'. The desired output would be '[][mohan][][rao][?][nagam]'.

Where do i need to change coding ?

Thank you for your help!

[Former Member]

Hi Clemens Leider ,

Now paste this code...modified..solves ur problem..

data:

w_string(50) type C value 'ram[mohan]hh[rao]k[nagam]',

w_len type i,

w_cnt type i.

data:

w_output type string ,

w_temp type string.

data:

itab like standard table of w_string with header line.

w_len = strlen( w_string ).

do w_len times.

w_cnt = sy-index - 1.

if w_string+w_cnt(1) eq '['.

w_string+w_cnt(1) = ']'.

endif.

enddo.

split w_string at ']' into table itab.

loop at itab .

w_cnt = sy-tabix mod 2.

if w_cnt eq 1.

<b>w_len = strlen( itab ).

if w_len = 1.

w_temp = '[?]'.

else.

w_temp = '[*]'.

endif.</b>

else.

w_temp = itab.

endif.

concatenate w_output w_temp into w_output.

clear w_temp.

endloop.

write: w_output.

[former_member5350]

Thank you very much, now it works as i wanted it to, but one last question remains - how to use this logic with a string rather than type c ?

[Former Member]

Ram's code works for works for strings as well.

[former_member5350]

Hi, i am afraid it doesnt because i get an error stating that at writing positions at fields of type string or xstring no offsetlenght is allowed - it refers to:

if w_string+w_cnt(1) eq '['.

w_string+w_cnt(1) = ']'.

endif.

can you help me ?

[Former Member]

Hi,

not to hijack the thread but see if this helps: Modified version of Ram's code

REPORT ZTEST.

data:
w_string type string value 'ram[mohan]hh[rao]k[nagam]',
*w_string(50) type c value 'ram[mohan]hh[rao]k[nagam]',
w_len type i,
w_cnt type i.

data:
w_output type string ,
w_temp type string.

data:
itab like standard table of w_string with header line.


w_len = strlen( w_string ).

*do w_len times.
*w_cnt = sy-index - 1.
*
*if w_string+w_cnt(1) eq '['.
*w_string+w_cnt(1) = ']'.
*endif.
*
*enddo.
translate w_string using '[]'.
split w_string at ']' into table itab.

loop at itab .
w_cnt = sy-tabix mod 2.

if w_cnt eq 1.

w_len = strlen( itab ).
if w_len = 1.
w_temp = '[?]'.
else.
w_temp = '[*]'.
endif.
else.
w_temp = itab.
endif.

concatenate w_output w_temp into w_output.
clear w_temp.
endloop.

write: w_output.

[Former Member]

Hi ,

Modified <b>...works with string</b> also...

data:

w_string type string value 'ram[mohan]hh[rao]kkk[ram]h[nagam]',

w_len type i,

w_cnt type i.

data:

w_output type string ,

w_temp type string.

data:

itab like standard table of w_string with header line.

translate w_string to upper case.

w_len = strlen( w_string ).

do w_len times.

w_cnt = sy-index - 1.

if w_string+w_cnt(1) na sy-abcde.

concatenate w_temp ']' into w_temp.

else.

concatenate w_temp w_string+w_cnt(1) into w_temp.

endif.

enddo.

w_string = w_temp.

split w_string at ']' into table itab.

loop at itab .

w_cnt = sy-tabix mod 2.

if w_cnt eq 1.

w_len = strlen( itab ).

if w_len = 1.

w_temp = '[?]'.

else.

w_temp = '[*]'.

endif.

else.

w_temp = itab.

endif.

concatenate w_output w_temp into w_output.

clear w_temp.

endloop.

write: w_output.

[Former Member]

Hi,

Following is the code in addition to what all has been said by others

data : string(20),

string1(20),

string2(20),

string3(20),

string4(20),

del value '[',

del2 value ']'.

string = 'M[ey]er'.

split string at del into string1 string2.

*write / string1.

split string2 at del2 into string3 string4.

*write / string4.

concatenate string1 string4 into string1.

write / string1. "Gives you Mer

[former_member5350]

Hi does this function also respect that there can be more than one [xyz] in a string ?

[Former Member]

Try this:

REPORT ZTEST.

data: v_name type string value 'Me[yjghje]r'.
data: str1 type string.
data: str2 type string.
data: dummy type string.

split v_name at '[' into dummy str1.
split str1 at ']' into str2 dummy.
concatenate '[' str2 ']' into str2.
replace all occurrences of str2  in v_name with  ''.
write:/ v_name.