Scanning and Trimming

[Ju5]

I am trying to use a query created database file as input for a RPG program. The database file has the following fields:
Field length
Number 7
Company 40
Country 15
Name 164

Name is a concatenation of 4 fields.

I defined the file in the RPG program as:

DFile DS
D Num 7
D Co 40
D Cty 15
D Nam 164

What I need to do is to search the Nam field for the characters 'ATTN:' or 'CONTACT:' and remove everything that preceds the ':'.

I think the SCAN command will be able to help me search the field. I've seen examples written in Free-form but not the standard form. Does anyone know how it should look like?

From what I read of the TRIM command it only removes leading/trailing blanks. Should I just Scan for the position of ATTN: and use it as the start of a SUBST?

 

[Ju5]

Is there a way to SCAN again starting from the position of the last SCAN performed?

Ex.

Name field contains:
This street, that country ;Attn:Mr.x/Mr.Y ;

I performed a scan for both ATTN and ; to get the length of a substring. In the example above, since the first ';' comes before 'ATTN' then the substring length is < 0. Is there a way to SCAN again so that I can get the position of the second ';'?

 

[SMURF75]

Set third parameter in scan to set start position.

- - - - - - - - - - - - - - - - - -
Im three apples high, Im blue, and i most certainly like that cold beer that should be every mans right after a hard days work!

 

[flapeyre]

Why can't you do this in free form? The %REPLACE BIF was made for this.

     D Valid_Chars     C                   '''-/ ABCDEFGHIJKLMNOPQRSTUVWXYZ0123-
     D                                     456789abcdefghijklmnopqrstuvwxyz&#'  
     D Pos             S              5U 0 Inz 
     D Nullstring      S              1A   Varying Inz('')

      /Free

         // Get rid of all invalid characters in CUSAD1 field.
         Dou Pos = 0;                                           
           Pos = %CHECK(Valid_Chars:Cusad1);                    
           If Pos > 0;                                          
             Cusad1 = %REPLACE(Nullstring:Cusad1:Pos:1);        
           Endif;                                               
         Enddo;

Solum potestis prohibere ignes silvarum.

 

[arrow483]

You want %scan or %replace, as suggested. However, you are wrong about %trim. It does only trim beginning/ending chars, but can trim ANY chars, not just blanks.
Also, if you have free-form and want to code in fixed, it pretty easy to convert back. (Can't think of a good reason to do so). For example flapeyre's code would look like this:

     C                   Dou       Pos = 0                                           
     C                   Eval      Pos = %CHECK(Valid_Chars:Cusad1)                    
     C                   If        Pos > 0                                          
     C                   Eval      Cusad1 = %REPLACE(Nullstring:Cusad1:Pos:1)        
     C                   Endif                                               
     C                   Enddo

 

[jmanj]

I usually do SUBSTR:

IF substr(nam:1,4) = "ATTN"
OR substr(nam:1,7) = "REPLACE"
DO.....

 

[jmanj]

Sorry, Didn't quite read the entire scope of your needs.

I usually use SUBSTR in conjuction with a do loop. Positioning
should all be a variable.

Check for the trailing spaces to give the length of the
name field to scan/test then subtract the length of your test variable "ATTN" and "REPLACE" this will give you the last position before you exit your do loop.

 

[Mercury2]

It does only trim beginning/ending chars, but can trim ANY chars, not just blanks.

Sorry Arrow but this statement can lead to put a wrong interpretation on the TRIM or %TRIM function.
- In SQL the TRIM function removes blanks or another specified character from the end or beginning of a string expression.
- In RPG IV the %TRIM function returns the given string less any leading and trailing blanks and solely blanks.

 

[arrow483]

>>In RPG IV the %TRIM function returns the given string less any leading and trailing blanks and solely blanks <<

Not true anymore.

The second parm on the %trim, %trimL, %trimR identifies the characters to be trimmed. Esp useful for dumping things like $ and * , for example.

NewField = %trim(***$5.67***:'*$') will trim off all the *'s and the $, but NOT any blanks (because blanks not in parm 2). Without the 2nd parm, it trims only blanks.

 

[flapeyre]

Good to know. However, that second parameter for %trim doesnt appear to be available at V5R3 and below. On those earlier releases, you can always use my little routine above (it worked on a V5R3 machine compiling to a previous release of V5R1).

Solum potestis prohibere ignes silvarum.

 

[Mercury2]

I didn't know that and learnt something new today. I have to say that I work on a V5R2 box and the second parm on the %TRIM doesn't exist yet.