|
View previous topic :: View next topic
|
| Author |
Message |
AlexSalas95
New User
Joined: 18 Mar 2024
Posts: 10
Location: United States
|
|
|
|
I've written a program that uses a GDG base as input. The GDG base may contain several generations. The input is sorted internally within the COBOL program, as opposed to using SyncSort/DFSort, due to processing requirements.
The issue I'm having is when I sort the file, it's also actually writing this sorted data into one of the generations. In my testing, it's generally the last generation but I believe it's also written to the initial generation also.
I'm fairly confident I know what my error is, but I'm not exactly sure what would be the best solution. See the incomplete COBOL code below;
| Code: |
IDENTIFICATION DIVISION.
PROGRAM-ID. ATEST.
******************************************************************
ENVIRONMENT DIVISION.
******************************************************************
CONFIGURATION SECTION.
******************************************************************
INPUT-OUTPUT SECTION.
******************************************************************
FILE-CONTROL.
******************************************************************
SELECT INFILE ASSIGN TO INFILE
FILE STATUS IS INFILE-STATUS.
SELECT SORT-INFILE ASSIGN TO SORT-INFILE.
SELECT OUTFILE1 ASSIGN TO OUTFILE1
FILE STATUS IS OUTFILE1-STATUS.
******************************************************************
DATA DIVISION.
******************************************************************
FILE SECTION.
******************************************************************
FD INFILE
BLOCK CONTAINS 0 RECORDS
LABEL RECORDS ARE STANDARD.
01 INFILE-RECORD.
02 IN-FIELD PIC X(80).
SD SORT-INFILE.
01 INFILE-RECORD-S.
02 IN-FIELD-S PIC X(80).
FD OUTFILE1
BLOCK CONTAINS 0 RECORDS
LABEL RECORDS ARE STANDARD.
01 OUTFILE1-RECORD PIC X(097).
******************************************************************
PROCEDURE DIVISION.
******************************************************************
SORT SORT-INFILE
ON ASCENDING KEY IN-FIELD-S
USING INFILE
GIVING INFILE.
|
I'm 99% it's when the last line shown,
executes.
For a visual, assume input data as follows;
input gdg0001v00 contains 1 record; ALPHA
input gdg0002v00 contains 1 record; BRAVO
input gdg0003v00 contains 1 record; CHARLIE
After performing the sort (alphabetically, ascending for this example), the data is as follows;
input gdg0001v00 contains 1 record; ALPHA
input gdg0002v00 contains 1 record; BRAVO
input gdg0003v00 contains 3 record; CHARLIE BRAVO ALPHA
How could I sort the input internally without altering the actual input gdgs? |
|
| Back to top |
|
 |
Phrzby Phil
Senior Member
Joined: 31 Oct 2006
Posts: 1050
Location: Richmond, Virginia
|
|
|
|
I never used the COBOL sort, but I find two things here confusing.
1.
| Quote: |
| ...but I believe it's also written to the initial generation also |
Yet your example shows only the last generation modified
2. Your sort says ASCENDING, yet your last generation's records are: CHARLIE BRAVO ALPHA |
|
| Back to top |
|
|
Pete Wilson
Active Member
Joined: 31 Dec 2009
Posts: 592
Location: London
|
|
|
|
| Seems a bizarre thing to be doing. Can you include the JCL used, |
|
| Back to top |
|
|
AlexSalas95
New User
Joined: 18 Mar 2024
Posts: 10
Location: United States
|
|
|
|
@Phrzby Phil
Sorry I should've made that clearer. It seems that when I sort ascending, it puts the records in the initial GDG. Descending puts it in the last GDG. In the example I gave, it should've been descending which would've resulted in the data shown
@Pete Wilson
Yeah, I may end up doing this another way. I think this way would've been fine had I not been using a GDG base as input. I'll drop the jcl in here anyway;
| Code: |
//S001 EXEC PGM=ATEST,COND=(4,LT),REGION=4M
//STEPLIB DD DSN=SYS2.TEST,DISP=SHR
//SYSOUT DD SYSOUT=*
//SYSPRINT DD SYSOUT=*
//* \/ THIS IS A GDG BASE
//INFILE DD DSN=USERNAME.SEQ.TEST.GDG,DISP=SHR
//OUTFILE1 DD DSN=USERNAME.SEQ.NEW.OUT,
// DISP=(NEW,CATLG,DELETE),UNIT=DISK,
// BUFNO=50,DATACLAS=F097,
// SPACE=(CYL,(1,1),RLSE) |
|
|
| Back to top |
|
|
Pete Wilson
Active Member
Joined: 31 Dec 2009
Posts: 592
Location: London
|
|
|
|
| Based on the JCL OUTFILE1 is a completely separate dataset and not a GDG, so how you end up with all the INFILE1 GDG generations combined into generation G0003V00 of INFILEI really don't know. The program also seems to refer to INFILE rather than INFILE1 as well. |
|
| Back to top |
|
|
AlexSalas95
New User
Joined: 18 Mar 2024
Posts: 10
Location: United States
|
|
|
|
For progeny, I wouldn't recommend doing this unless you have to. And if you think you have to, there's probably another, better way
However, a solution I found is to just write (or "give") the sort to a separate output file as opposed to an input. When I originally wrote the code, it felt wrong to write to my input anyway, but it would've been fine had it not been a GDG base
The COBOL sort to do this is very simple;
| Code: |
SORT SORT-INFILE
ON ASCENDING KEY IN-FIELD-S
USING INFILE
GIVING OUTFILE1. |
I generally use a sort utility to sort my input, but it would've been a little tricky for this program in particular. Also, the input is only a few hundred, maybe a thousand rows per run, so performance wasn't really a concern. |
|
| Back to top |
|
|
AlexSalas95
New User
Joined: 18 Mar 2024
Posts: 10
Location: United States
|
|
|
|
@Pete Wilson
The "giving" command in the internal sort function of the program writes the output to the file specified. The file does not need to be opened for output (or opened at all) for it do this, allowing you to output to an input file
Sketchy, but I think for most applications it's safe since it's only sorting the file and not changing record lengths or anything |
|
| Back to top |
|
|
Phrzby Phil
Senior Member
Joined: 31 Oct 2006
Posts: 1050
Location: Richmond, Virginia
|
|
|
|
My preference was always a separate JCL Sort: COBOL-Program-A, Sort, COBOL-Program-B.
I found this kept my logic simpler, plus very few other programmers I worked with used COBOL Sort, so much safer. |
|
| Back to top |
|
|
|
|
