Hi,

Let said I have below field value:

NO. | Description
1 | help34to**solve proble(m
2 | ho23w yo44u do99ing toda y?
3 | &th@is is (very) n iceay

I want to turn into this:

NO. | Description_2
1 | helptosolveproblem
2 | howyoudoingtoday
3 | thisisveryniceday

those un-wanted characters may be all over places, I checked STRREP but I don't think length would help. any idea?

This message has been edited. Last edited by: nox,

You can define a custom function or simply use REPLACE
But since it seems that you have a bunch of character to replace, may be complex because you will have to provide all special characters to replace

DEFINE FILE abc
newField /A250 = REPLACE(REPLACE(REPLACE(REPLACE(REPLACE(REPLACE(REPLACE (oldField, '1', ''), '2', ''), '3', ''), '4', ''), '*', ''), '(', ''), ')', '');
END

Thanks MartinY, this is an idea, but is there a better way to extract only 'A-Z' and 'a-z' only?

Out of the box IBI functions are to remove/replace characters not to keep desired ones.

Maybe you can accomplish that using SQL passthrough

MartinY thank you for your information. I did try out using SQL function to do this and it work well, but do wish IBI can have this feature on it own, so don't have too rely on others.

I cannot tell you
But personally, I would say no. I don't think that IBI will try to implement all possible function/feature that SQL can do.
Having the ability to perform SQL passthrough is the answer to be able to accomplish almost anything from SQL : especially that depending on which SQL you connect to, the functions differs.

You will need to open a case and ask the question. It's only from there that you will get the right answer.

Not sure if it would help but look at the REGEX function - link

T

Here is a C user written function: first, copy the following code into nospchar.c:

#include <stdio.h>
#include <string.h>
#include <memory.h>

const unsigned char nospasc[256] = {
0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
0,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,
1,1,1,1,1,1,1,1,1,1,1,0,0,0,0,0,
0,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,
1,1,1,1,1,1,1,1,1,1,1,0,0,0,0,0,
0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
};
/*-------------------------------------------------------------------*/
/* xnospchar: return input text without special characters           */
/*-------------------------------------------------------------------*/
                                        /*                           */
int nospchar(unsigned char *input, double *dlen, unsigned char *output)
{
int len = (int)*dlen ;
unsigned char *in,*out,*end ;
if(len>256 || len<0) len = 0;

end = input + len;
memset(output,' ',len);                 /* blank out output          */
for(in=input,out=output; in<end; in++) {
  if(nospasc[*in]) *out++ = *in;
  }
return(0);
}

/* this commented out main will printout the nospchar text from the
   command line.
int main(int argc, char **argv)
{
unsigned char outc[256];
int len ;
double dlen;

memset(outc,0,sizeof(outc));
printf("argc: %d %s\n",argc,argv[1]);
if(argc != 2) { printf("usage: nospchar <string>\n"); return(0); }
len = strlen(argv[1]);
dlen = len ;
nospchar((unsigned char *)argv[1],&dlen,outc);
printf("%s\n",outc);
return(0);
}
*/

/* this commented out main prints out the nospasc array above for
   upper and lower case latin1 characters
int main(int argc, char **argv)
{
unsigned char lower[26] = "abcdefghijklmnopqrstuvwxyz" ;
unsigned char upper[26] = "ABCDEFGHIJKLMNOPQRSTUVWXYZ" ;
int i,j;
memset(nospasc,0,sizeof(nospasc));
for(i=0; i<26; i++) nospasc[lower[i]] = 1;
for(i=0; i<26; i++) nospasc[upper[i]] = 1;

for(i=0; i<16; i++) {
  for(j=0; j<16; j++) {
    printf("%d,",nospasc[i*16+j]);
    }
  printf("\n");
  }
return(0);
} */

 

Now, unfortunately you must read the documentation for your platform on how to make user functions. Usually, the IBI provided gencpgm.sh will create a compiled and linked program.
For instance, under linux you can issue:

  
gencpgm.sh -g -e -v -m cpgm nospchar.c
 

Remember to issue the export IBICPG environment variable:

export IBICPG= [directory of user functions]

e.g. export IBICPG=/mypath/usrfuncpath

Here is a focexec that tests nospchar:

  
DEFINE FILE CAR
MYC/A15 = EDIT(COUNTRY,'9%lh9+*999999$9') ;
MYD/A15 = NOSPCHAR(MYC,15,MYD) ;
END
TABLE FILE CAR
PRINT COUNTRY MYC MYD
END
-RUN


So, we execute it: 

EX TEST

To get: 
 PAGE     1
 
 
 COUNTRY     MYC              MYD                                              
 -------     ---              ---                                              
 ENGLAND     E%lhN+*GLAND     ElhNGLAND      
 JAPAN       J%lhA+*PAN       JlhAPAN        
 ITALY       I%lhT+*ALY       IlhTALY        
 W GERMANY   W%lh +*GERMAN    WlhGERMAN      
 FRANCE      F%lhR+*ANCE      FlhRANCE       
 
 

This message has been edited. Last edited by: Edward Wolfgram,

quote:

...but do wish IBI can have this feature on it own, so don't have too rely on others.

The WF function, DB_EXPR, will allow you to use native SQL functionality, within the WF reporting language.

Example running against the WF Retail Lite MS SQL Server database:

DEFINE FILE WF_RETAIL_STORE
 AREARANGESTFT/A20V = DB_EXPR(REPLACE(TRANSLATE("AREA_RANGE_SQ_FT", '0123456789 .-:,', '@@@@@@@@@@@@@@@'), '@', '') );
 AREARANGESTM/A20V  = DB_EXPR(REPLACE(TRANSLATE("AREA_RANGE_SQ_M",  '0123456789 .-:,', '@@@@@@@@@@@@@@@'), '@', '') );
END
-*
TABLE FILE WF_RETAIL_STORE
"'Keep' Only 'A-Z' Characters In a String."
PRINT AREA_RANGE_SQ_FT
      AREA_RANGE_SQ_M
      AREARANGESTFT
      AREARANGESTM
-*
ON TABLE SET STYLE *
  INCLUDE = IBFS:/EDA/EDASERVE/_EDAHOME/ETC/jellybean_combo.sty, $
  TYPE=DATA, JUSTIFY=CENTER,$
 ENDSTYLE
-*
IF READLIMIT EQ 10
IF AREA_RANGE_SQ_FT IS-NOT MISSING
END 

In the SQL trace you can see the DB_EXPR value, within the generated SQL:

SELECT
TOP 10 (REPLACE(TRANSLATE(T1."AREA_RANGE_SQ_FT", '0123456789
.-:,', '@@@@@@@@@@@@@@@'), '@', '') ),
(REPLACE(TRANSLATE(T1."AREA_RANGE_SQ_M",  '0123456789 .-:,',
'@@@@@@@@@@@@@@@'), '@', '') ),
T1."AREA_RANGE_SQ_FT",
T1."AREA_RANGE_SQ_M"
FROM
wrd_wf_retail_store T1
WHERE
(T1."AREA_RANGE_SQ_FT" IS NOT NULL); 

This message has been edited. Last edited by: David Briars,

Hi,
didn't thinks EDIT could receive a field as second argument but it does , funny

DEFINE FILE CAR
T1/A100 WITH CAR = IF CAR EQ 'AUDI' THEN 'HEPto34**TOSOLCE'
ELSE IF CAR EQ 'BMW' THEN 'HDW FFF ? DDD DD' ELSE 'HEPto34';
P1/A100 = PATTERN(100,T1,'A100');
R1/A100 = STRREP(100,P1,1,'9',1,'$',100,R1);
R2/A100 = STRREP(100,R1,1,'A',1,'9',100,R2);
R2/A100 = STRREP(100,R2,1,'a',1,'9',100,R2);
R2/A100 = STRREP(100,R2,1,'*',1,'$',100,R2);
R2/A100 = STRREP(100,R2,1,' ',1,'$',100,R2);
R2/A100 = STRREP(100,R2,1,'?',1,'$',100,R2);
F1/A100 = EDIT(T1,R2);
END

TABLE FILE CAR
PRINT T1 P1 R1 R2 F1
CAR
END