I am using NetCOBOL for .NET to extend a COBOL application. When I read through and input file I am doing a simple comparison on a numeric field and on certain records I get an INVALID VALUE SPECIFIED
error.
The field is INV-MST-ONHAND as defined:
02 INV-MST-LOCATION OCCURS 3.
04 INV-MST-ONHAND PIC S9(7)V999.
04 INV-MST-SALES PIC S9(6) OCCURS 12.
04 INV-MST-PTDSALE PIC S9(6)V999.
The comparison is in this line of code
IF INV-MST-ONHAND(1) > 0 MOVE INV-MST-ONHAND(1) TO WS-ITEM-ONHAND.
It only happens on some records. The data file is indexed so it's not easy to parse by hand. I am looking for a way to fix it or even to skip over these but I don't know how to catch and handle this error condition.
Here is the full record layout:
000381 FD INVMAST
000382 RECORD 845 CHARACTERS.
000383 01 INV-MST-REC.
000384 02 INV-MST-MCS PIC X(10).
000385 02 INV-MST-KEY.
000386 04 INV-MST-ITEM-NO PIC X(15)
000387 02 INV-MST-ALT1-KEY.
000389 04 INV-MST-PLU PIC 9(13).
000390 02 INV-MST-ALT2-KEY.
000391 04 INV-MST-WORDS1 PIC X(10).
000392 02 INV-MST-ALT3-KEY.
000393 04 INV-MST-WORDS2 PIC X(10).
000394 02 INV-MST-ALT4-KEY PIC X(10).
000395 02 INV-MST-ALT5-KEY.
000396 04 INV-MST-CATEGORY PIC X(10).
000399 02 INV-MST-UPC-KEYS.
000400 04 INV-MST-UPC01-KEY PIC X(15).
000401 04 INV-MST-UPC02-KEY PIC X(15).
000402 04 INV-MST-UPC03-KEY PIC X(15).
000403 04 INV-MST-UPC04-KEY PIC X(15).
000404 04 INV-MST-UPC05-KEY PIC X(15).
000405 04 INV-MST-UPC06-KEY PIC X(15).
000406 04 INV-MST-UPC07-KEY PIC X(15).
000407 04 INV-MST-UPC08-KEY PIC X(15).
000408 04 INV-MST-UPC09-KEY PIC X(15).
000409 04 INV-MST-UPC10-KEY PIC X(15).
000410 04 INV-MST-UPC11-KEY PIC X(15).
000411 04 INV-MST-UPC12-KEY PIC X(15).
000412 02 INV-MST-ITEM-NAME.
000454 04 INV-MST-NAME1 PIC X(20).
000455 04 INV-MST-NAME2 PIC X(20).
000456 04 INV-MST-NAME3 PIC X(20).
000457 02 INV-MST-SCRAP PIC 9(6)V99.
000458 02 INV-MST-ENV-LEVY PIC 9(3)V99.
000459 02 INV-MST-UORD PIC XXX.
000460 02 INV-MST-USTOCK PIC XXX.
000461 02 INV-MST-UUNO PIC 99999.
000462 02 INV-MST-VEND1 PIC 9999.
000463 02 INV-MST-VEND2 PIC 9999.
000464 02 INV-MST-VEND3 PIC 9999.
000465 02 INV-MST-ONORDER PIC 9(6).
000466 02 INV-MST-COMMIT PIC 9(6).
000467 02 INV-MST-REORD PIC 9(5).
000468 02 INV-MST-MINORD PIC 9(5).
000469 02 INV-MST-BASIS PIC 99.
000470 02 INV-MST-BIN PIC X(6).
000471 02 INV-MST-GL PIC 9(4).
000472 02 INV-MST-DEPT PIC 999.
000473 02 INV-MST-ALTITEM1 PIC X(15).
000474 02 INV-MST-ALTITEM2 PIC X(15).
000476 02 INV-MST-COST1 PIC 9(6)V99.
000477 02 INV-MST-COST2 PIC 9(6)V99.
000478 02 INV-MST-COST3 PIC 9(6)V99.
000479 02 INV-MST-COST4 PIC 9(6)V99.
000480 02 INV-MST-SELLPRICES.
000481 04 INV-MST-SELL1 PIC 9(6)V99.
000482 04 INV-MST-SELL2 PIC 9(6)V99.
000483 04 INV-MST-SELL3 PIC 9(6)V99.
000484 04 INV-MST-SELL4 PIC 9(6)V99.
000485 04 INV-MST-SELL5 PIC 9(6)V99.
000486 02 FILLER REDEFINES INV-MST-SELLPRICES.
000487 04 INV-MST-SELL PIC 9(6)V99 OCCURS 5.
000488 02 INV-MST-SELLPRICE-FACTORS.
000489 04 INV-FACTOR OCCURS 5.
000490 06 INV-PRC-FACTOR PIC 9V99.
000491 06 INV-COST-BASE PIC X.
000492 02 INV-MST-SALE-DECPT PIC 9.
000493 02 INV-MST-STOCK-DECPT PIC 9.
000494 02 INV-MST-PRICE-UNITS PIC 9999.
000495 02 INV-MST-ALTERNATE-UNITS.
000496 04 INV-MST-ALT-DESC PIC X(10).
000497 04 INV-MST-ALT-ABBREV PIC XX.
000498 04 INV-MST-ALT-FACTOR PIC 9(3)V9(4).
000499 04 INV-MST-ALT-SALE-FLAG PIC X.
000500 04 INV-MST-ALT-ORDER-FLAG PIC X.
000501 02 INV-MST-TAX1 PIC X.
000502 02 INV-MST-TAX2 PIC X.
000503 02 INV-MST-LOCATION OCCURS 3.
000504 04 INV-MST-ONHAND PIC S9(7)V999.
000505 04 INV-MST-SALES PIC S9(6) OCCURS 12.
000506 04 INV-MST-PTDSALE PIC S9(6)V999.
000507 02 INV-MST-FIL PIC X(30).
When I check the value of INV-MST-ONHAND(1)
when it fails, Visual Studio debugger shows a value of +0000000.000 and the condition tests as true using Quick Watch.
You are using a compiler option,
CHECK
, with sub-option(NUMERIC)
or(ALL)
.This is producing the
JMP0828I-U
message. TheU
means it will end execution. Somewhere you could tell it to instead produceE
as a suffix to the message, in which case you'd get the message but execution would continue.The problem is that you data does not "conform to PICture". You field is defined as signed (the
S
in your PICture string), but the overpunch which would indicate a positive or negative value is not present.You mentioned that you created the data. What I suspect is that where you created the data you accidentally described that field as unsigned (no
S
in the PICture string) and then, for expedience, did a group-move to the record, allowing an unsigned value to get into your signed field.The CHECK(NUMERIC) or CHECK(ALL) spotted this when you attempted the IF-test, and produced the message (and ended your program).
It is interesting that your
NUMERIC
test was true. There is probably a compiler-option to describe what is valid in the NUMERIC test. You'd need to locate this, and ensure that only positive and negative, and not unsigned, are valid for a signed field.A couple of lessons: the message reference is important, and allows you to locate more information about the message - for your future searches, exclude the suffix, as the description in the documentation does not describe the suffixes separately; avoid group MOVEs - or take great care with them - a missing sign, or the wrong length or number of decimal places can give you subtle errors, its not worth the saving of coding vs time chasing the subtleties.
With your debugger, you should see if there is some way to see the actual source value, not the "edited" value that you showed in your question. Seeing the actual source would have aided you earlier.
Non-COMPUTATIONAL numerics are "zoned". One digit per byte, preceded by the "zone" value. For an unsigned value, all digits have zoned. For a signed value the low-order byte has the "overpunch" to indicate positive or negative instead.
An unsigned value is always treated as positive.
COBOL to the 1985 Standard has no mechanism to "catch" errors caused by non-numeric values. Your documentation will indicate whether you have any way to do that with your NetCOBOL run-time.