|
View previous topic :: View next topic
|
| Author |
Message |
sandeep kumar302
New User
.jpg)
Joined: 14 Mar 2012
Posts: 81
Location: India
|
|
|
|
Hi,
My requirement is-
Valid characters only for detail records positions 31-62 (32 characters in length).Only the following characters should be allowed.
| Code: |
a. "a" THRU "z”
b. "A" THRU "Z"
c. "0" THRU "9"
d. " " or blank fill
e. "."
f. "&" |
All other characters should be replaced with a space character, for example “M’Brien” should be “ M Brien” or “ABC PVT Ltd <Accnt a/c>” should be “ABC PVT Ltd Accnt a c”
Can anyone please help me on this?
Code'd |
|
| Back to top |
|
 |
Bill Woodger
Moderator Emeritus
Joined: 09 Mar 2011
Posts: 7309
Location: Inside the Matrix
|
|
|
|
In your sample output, you are doing more than just replacing characters by blank.
If that's a typo, an INSPECT ... CONVERTING ... will do what you want. You just list all the characters you don't want, and an equivalent number of spaces for the result. |
|
| Back to top |
|
|
sandeep kumar302
New User
Joined: 14 Mar 2012
Posts: 81
Location: India
|
|
|
|
Thanks Bill,
I tried this solution like below:
| Code: |
INSPECT ACCOUNT-TITLE CONVERTING
"!@#$%^*()`~,<>/?;:'"[{]}\|" TO
" ".
|
but it gives me some error as "" is confusing to the compiler.
| Code: |
3538 INSPECT ACCOUNT-TITLE CONVERTING
3539 "!@#$%^*()`~,<>/?;:'"[{]}\|" TO
* 83-S********************************** ( 0)**
** Reserved word missing - TO expected.
* 4-S*********************************** ( 62)**
** Illegal character
* 4-S************************************* ( 62)**
** Illegal character
* 4-S************************************** ( 62)**
** Illegal character
* 4-S*************************************** ( 62)**
** Illegal character
*1004-E****************************************************************( 62)**
** Continuation character expected. End of literal assumed.
3540 " ".
|
but, when i remove the literal ", the program is compiled successfully like below:
| Code: |
INSPECT AUD-ACCOUNT-TITLE CONVERTING
"!@#$%^*()`~,<>/?;:'[{]}\|" TO
" ".
INSPECT AUD-ACCOUNT-TITLE REPLACING ALL '"' BY " ".
|
Is there any way i can get the result in single INSPECT as you suggested? |
|
| Back to top |
|
|
Bill Woodger
Moderator Emeritus
Joined: 09 Mar 2011
Posts: 7309
Location: Inside the Matrix
|
|
|
|
If you have a quote in a literal, you can bound with apostrophes. If you have an apostrophe in a literal, you can bound with quotes.
If you have both apostrophe and quote in a literal...
You can define in WORKING-STORAGE, and use that as the source of the COVERTING.
You can use a hexadecimal literal.
Slight performance penalty with the former, quite large obfuscation penalty with the latter. |
|
| Back to top |
|
|
sandeep kumar302
New User
Joined: 14 Mar 2012
Posts: 81
Location: India
|
|
|
|
Thanks Bill for the prompt reply.
I already tried bounding with apostrophe like below:
| Code: |
INSPECT AUD-ACCOUNT-TITLE CONVERTING
"!@#$%^*()`~,<>/?;:'[{]}\|'"'" TO
" ".
INSPECT AUD-ACCOUNT-TITLE REPLACING ALL '"' BY " ".
|
But this gave me a compilation error. I am using microfocus cobol.
Actually my motive is not to have anything from the literals
!@#$%^*()`~,<>/?;:'[{]}\| and " also should not be present like abc"xyz..anything like this must be replaced by abc xyz |
|
| Back to top |
|
|
sandeep kumar302
New User
Joined: 14 Mar 2012
Posts: 81
Location: India
|
|
|
|
Though i can try the other option which you suggested:
| Quote: |
You can define in WORKING-STORAGE, and use that as the source of the COVERTING.
|
|
|
| Back to top |
|
|
Bill Woodger
Moderator Emeritus
Joined: 09 Mar 2011
Posts: 7309
Location: Inside the Matrix
|
|
|
|
Sorry for not being clear. With both apostrophe and quote in the literal, you can't do it with an ordinary literal.
It has to be defined as data, you'll have to define the separate parts separately.
The only literal you can use is a hex literal, and that would be very obscure and typo-prone. |
|
| Back to top |
|
|
Bill O'Boyle
CICS Moderator
Joined: 14 Jan 2008
Posts: 2501
Location: Atlanta, Georgia, USA
|
|
|
|
| Please list the characters (in hex) that you need to convert to SPACE. Then, we'll go from there.... |
|
| Back to top |
|
|
Bill O'Boyle
CICS Moderator
Joined: 14 Jan 2008
Posts: 2501
Location: Atlanta, Georgia, USA
|
|
|
|
Maybe this will do the trick (desk checked only) -
| Code: |
03 WS-SUB PIC 9(08) BINARY.
03 WS-SUB-X REDEFINES WS-SUB
PIC X(04).
03 WS-FWORD PIC 9(08) BINARY.
03 WS-FWORD-X REDEFINES WS-FWORD
PIC X(04).
03 AUD-ACCOUNT-TITLE PIC X(79).
03 WS-XLATE-TBL-IN PIC X(256).
03 WS-XLATE-TBL-OUT PIC X(256).
PERFORM VARYING WS-FWORD FROM 1 BY 1
UNTIL WS-FWORD > LENGTH OF WS-XLATE-TBL-IN
COMPUTE WS-SUB = (WS-FWORD - 1)
MOVE WS-SUB-X (4:) TO WS-XLATE-TBL-IN (WS-FWORD:1)
END-PERFORM
*
**** BUILD THE 'TBL-OUT', MOVING LOWER-CASE AND UPPER-CASE LETTERS
**** LETTERS AS WELL AS NUMERICS 0-9. FOR EXAMPLE, 130 IS A LOWER
**** CASE 'b'. BUT, 130 IS THE TBL-OFFSET, WHERE YOU'LL FIND A
**** LOWER-CASE 'a'.
*
MOVE SPACES TO WS-XLATE-TBL-OUT
MOVE 130 TO WS-SUB
MOVE WS-XLATE-TBL-IN (WS-SUB:9)
TO WS-XLATE-TBL-OUT (WS-SUB:9)
MOVE 146 TO WS-SUB
MOVE WS-XLATE-TBL-IN (WS-SUB:9)
TO WS-XLATE-TBL-OUT (WS-SUB:9)
MOVE 163 TO WS-SUB
MOVE WS-XLATE-TBL-IN (WS-SUB:8)
TO WS-XLATE-TBL-OUT (WS-SUB:8)
MOVE 194 TO WS-SUB
MOVE WS-XLATE-TBL-IN (WS-SUB:9)
TO WS-XLATE-TBL-OUT (WS-SUB:9)
MOVE 210 TO WS-SUB
MOVE WS-XLATE-TBL-IN (WS-SUB:9)
TO WS-XLATE-TBL-OUT (WS-SUB:9)
MOVE 227 TO WS-SUB
MOVE WS-XLATE-TBL-IN (WS-SUB:8)
TO WS-XLATE-TBL-OUT (WS-SUB:8)
MOVE 241 TO WS-SUB
MOVE WS-XLATE-TBL-IN (WS-SUB:10)
TO WS-XLATE-TBL-OUT (WS-SUB:10)
INSPECT AUD-ACCOUNT-TITLE CONVERTING WS-XLATE-TBL-IN
TO WS-XLATE-TBL-OUT
* |
By initializing 'TBL-OUT' to SPACES and then, populating 'TBL-OUT' with the characters you'd like to keep, ensures the characters you don't want to keep will equal a SPACE in their respective slots in 'TBL-OUT'.
You may want to consider this as a CALLED sub-program.
HTH.... |
|
| Back to top |
|
|
RahulG31
Active User
Joined: 20 Dec 2014
Posts: 446
Location: USA
|
|
|
|
| If I understand this correctly then having an in-line Perform 32 times for each byte will do. We just have to check for Numeric, Alphabetic, '.' and '&'. I.e. If the byte Is Numeric Or Is Alphateic Or equal to '.' Or '&', then Continue Else move space to that byte. |
|
| Back to top |
|
|
Bill Woodger
Moderator Emeritus
Joined: 09 Mar 2011
Posts: 7309
Location: Inside the Matrix
|
|
|
|
RahulG31,
That'll be deeply slower than INSPECT. Until necessary (when TS/OP says "yes, I need to entirely remove some bytes") a manual loop is a poor way to do it vs INSPECT. |
|
| Back to top |
|
|
Rohit Umarjikar
Global Moderator
Joined: 21 Sep 2010
Posts: 3076
Location: NYC,USA
|
|
|
|
You can also perform a loop till last byte and check each byte and if you find (what you want to replace with space) then move space else move current value and end the loop. Is that what you want?
Ops.. It is the same as what RahulG31 suggested. We have similar in place but never got any substantial performance issues but technically it may be as per Bill. |
|
| Back to top |
|
|
prino
Senior Member
Joined: 07 Feb 2009
Posts: 1315
Location: Vilnius, Lithuania
|
|
|
|
| Sheesh, don't tell me that COBOL doesn't understand that '''' is a single apostrophe and """" a single quote... |
|
| Back to top |
|
|
Bill Woodger
Moderator Emeritus
Joined: 09 Mar 2011
Posts: 7309
Location: Inside the Matrix
|
|
|
|
The problem is that both apostrophe and quote are wanted in the same literal, and the doubling-up-to-give-one only works for the symbol which bounds the literal.
| Code: |
"""''" will give "''
'''""' will give '"" |
So you have to define the field as separate items for at least one of apostrophe/quote, and you can't use a literal (there's no literal concatenation in Enterprise COBOL).
If you get to PIC X VALUE "'" and PIC X VALUE '"', there's no real value in using PIC X VALUE """" or PIC X VALUE ''''.
So you can do it, but not within the same literal, and since you have a choice (these days) of what bounds a literal you just use the other one, I doubt you'll see the doubling-up used in code. |
|
| Back to top |
|
|
RahulG31
Active User
Joined: 20 Dec 2014
Posts: 446
Location: USA
|
|
|
|
@Bill: The problem I see with the use of INSPECT CONVERTING is that we are restricting ourselves to what we want to convert. To me, this is Not the requirement.
The requirement states what is Valid; Rest all is Invalid. And here, what we are trying to do with INSPECT is, telling what is Invalid and rest all is valid.
We have to make sure that we have all the symbols covered in INSPECT. As of now, I don't see '=, +, -' covered (there may possibly be more), which is error prone.
What TS/OP wants is (as I said earlier) Alphabetic, Numeric, '.' and '&' as Valid and Rest all is Invalid.
IS ALPHABETIC clause covers for capital and small letters as well as space.
IS NUMERIC covers 0-9.
. |
|
| Back to top |
|
|
Robert Sample
Global Moderator
Joined: 06 Jun 2008
Posts: 8700
Location: Dubuque, Iowa, USA
|
|
|
|
I looked at the original post and came up with this code (for testing I used a second PERFORM against INPUT-DATA-2 but that wouldn't be needed usually):
| Code: |
77 STR-POS PIC 9(02).
77 TEST-CHAR PIC X(01).
77 INPUT-DATA-1 PIC X(32)
VALUE "M'BRIEN".
77 INPUT-DATA-2 PIC X(32)
VALUE "ABC PVT Ltd <ACCNT A/C>".
77 OUTPUT-DATA PIC X(32).
PROCEDURE DIVISION.
0000-START.
PERFORM
VARYING STR-POS
FROM 1 BY 1
UNTIL STR-POS > 32
MOVE INPUT-DATA-1 (STR-POS : 1)
TO TEST-CHAR
IF TEST-CHAR ALPHABETIC
OR TEST-CHAR NUMERIC
OR TEST-CHAR = '.' OR '&'
MOVE INPUT-DATA-1 (STR-POS : 1)
TO OUTPUT-DATA (STR-POS : 1)
ELSE
MOVE SPACE
TO OUTPUT-DATA (STR-POS : 1)
END-IF
END-PERFORM.
DISPLAY INPUT-DATA-1.
DISPLAY OUTPUT-DATA.
DISPLAY ' ' .
|
which produces output of
| Code: |
M'BRIEN
M BRIEN
ABC PVT Ltd <ACCNT A/C>
ABC PVT Ltd ACCNT A C
|
If each character is valid, move it to the output variable character else move a space to that output character. This lets you avoid the whole QUOTE / APOST compiler option mess and ensuing complications. |
|
| Back to top |
|
|
Bill Woodger
Moderator Emeritus
Joined: 09 Mar 2011
Posts: 7309
Location: Inside the Matrix
|
|
|
|
Well, looking back I see the words Micro Focus COBOL (not spelled correctly).
Don't know if it is Micro Focus for a Mainframe target, or just Micro Focus. |
|
| Back to top |
|
|
Bill Woodger
Moderator Emeritus
Joined: 09 Mar 2011
Posts: 7309
Location: Inside the Matrix
|
|
|
|
Eek.
|
|
| Back to top |
|
|
|
|
