IBM MAINFRAME HELP & SUPPORT FORUMS

lanand_hps · Joined: 05 Dec 2007 Posts: 59 Location: chennai

Hi,
Can we sum fields with decimals?
The input file is like...

(column 3) (column 68)
003042193 0000009.08
003042193 0000009.08
003042193 0000009.08
007505628 0000001.73
007505628 0000001.73
007505628 0000001.73

I need to sum the amount fields which starts from 68th column and is of 10 digits including decimal. How do we do it?

I'm getting S0C7 if i use the control card mentioned below.
SORT FIELDS=(3,9,CH,A)
SUM FIELDS=(68,10,ZD)

Please comment.

Garry Carroll · Joined: 08 May 2006 Posts: 171 Location: Dublin, Ireland

You need to move the digits to a work-field, dropping the decimal. Then SUM the work-field and edit, if required to reinstate the decimal, to output field.

Where do you want to put the result of the SUM ?

Regards,
Garry

karthikr44 · Joined: 25 Aug 2007 Posts: 172 Location: Chennai

Hi,

You can use FL format for floating point numbers. That itself
u can use only 4, 8, or 16 bytes. i.e. u can use

SUM FIELDS=(11,16,FL)

But not

SUM FIELDS=(11,15,FL)

Regards
R KARTHIK

Garry Carroll · Joined: 08 May 2006 Posts: 171 Location: Dublin, Ireland

Quote: You can use FL format for floating point numbers.

Perhaps, but that's not what the O/P asked for. Floating-point is a totally different kettle of fish to what is described.

Garry.

karthikr44 · Joined: 25 Aug 2007 Posts: 172 Location: Chennai

Hi,

I posted because the reason for SOC7 abend is he is using ZD format
for numbers like 0002.3

Regards
R KARTHIK

Garry Carroll · Joined: 08 May 2006 Posts: 171 Location: Dublin, Ireland

Hi,

Yes, he is. but treating the numbers like 0002.3 as floating point won't solve his problem - he'd have to first change the format of the data to floating point.

Regards,
Garry.

Frank Yaeger · Joined: 15 Feb 2005 Posts: 4613 Location: San Jose, CA

Karthik,

FL is floating point format for numbers with a characteristic and fraction. It has nothing to do with values like 'ddddddd.dd' that the OP is asking about - you could have looked up FL format at:

http://publibz.boulder.ibm.com/cgi-bin/bookmgr_OS390/BOOKS/ICE1CA20/C.1?DT=20060615185603

lanand_hps,

You can use the technique discussed in the "Sum a number with a decimal point" Smart DFSORT Trick at:

www.ibm.com/systems/support/storage/software/sort/mvs/tricks/

lanand_hps · Joined: 05 Dec 2007 Posts: 59 Location: chennai

Thanks Frank.
I had created a jcl with two steps to sum the digits with decimal points. First step to remove decimal and treat that as a ZD. Second step to inser the decimal point in the appropriate postion.
FL didnot work.
But one more clarification.
Under what scenarios can we use PD?
For test purpose, i had 3 records in my postion with a packed decimal from 20th postion till 25th.

SORT FIELDS=(107,9,CH,A)
SUM FIELDS=(20,5,PD)

It is resulting in S0C7. Any details?
I checked the input file and it had proper values.

Craq Giegerich · Joined: 19 May 2007 Posts: 1001 Location: Virginia, USA

lanand_hps wrote: Thanks Frank.
I had created a jcl with two steps to sum the digits with decimal points. First step to remove decimal and treat that as a ZD. Second step to inser the decimal point in the appropriate postion.
FL didnot work.
But one more clarification.
Under what scenarios can we use PD?
For test purpose, i had 3 records in my postion with a packed decimal from 20th postion till 25th.

SORT FIELDS=(107,9,CH,A)
SUM FIELDS=(20,5,PD)

It is resulting in S0C7. Any details?
I checked the input file and it had proper values.

Show us the hex display of the records.

Frank Yaeger · Joined: 15 Feb 2005 Posts: 4613 Location: San Jose, CA

lanand_hps,

You can only use PD for valid packed decimal data which has the following form in hex:

dd...ds

where d is a digit (0-9) and s is a sign (usually C or F for + or D for -).

For example:

12345C is the PD equivalent of +12345
00030D is the PD equivalent of -30

If you try to SUM with PD for values that are NOT PD, you will get an 0C7 ABEND.

Note that PD values do NOT contain a decimal point internally.

Quote: For test purpose, i had 3 records in my postion with a packed decimal from 20th postion till 25th.

SORT FIELDS=(107,9,CH,A)
SUM FIELDS=(20,5,PD)

20-25 is 6 bytes, but you are using 20,5,PD instead of 20,6,PD. If you got an 0C7 ABEND, then the 5 bytes from 20-24 contain invalid PD data.

You can use the VERIFY operator of DFSORT's ICETOOL to display invalid PD values.

lanand_hps · Joined: 05 Dec 2007 Posts: 59 Location: chennai

Hi Craq,
This is how the file is now...

Code: decimal
------
000001 2003941297 07353LAD ��VOL-HERTZW
FFFFFFFFFF4FFFFFDCC000098EDD6CCDEEE44
200394129700735331400000C563085939600
--------------------------------------------
000002 2003941297 07354LAD VOL-HERTZW
FFFFFFFFFF4FFFFFDCC000013EDD6CCDEEE44
200394129700735431400007C563085939600
--------------------------------------------

lanand_hps · Joined: 05 Dec 2007 Posts: 59 Location: chennai

Sorry guys..
It is of 6 digits..as Frank pointed out.. 20th to 25th column.
One more doubt. ((Please spare me :) )

I have read in a lot of places that
Statement 1:
SORT FIELDS=COPY
SUM FIELDS=NONE will remove the duplicates.
Statement 2:
SORT FIELDS=NONE will also avoid duplicates.

When i just tested it, the duplicates are not getting removed with the first statement and statement 2 is throwing an error.

I guess it is a very basic doubt :( .. Kindly give your comments..

Craq Giegerich · Joined: 19 May 2007 Posts: 1001 Location: Virginia, USA

You either need to use 21,5,PD or 20,6,PD because the byte of the PD is in position 25 not 24.

Frank Yaeger · Joined: 15 Feb 2005 Posts: 4613 Location: San Jose, CA

Quote: Sorry guys..
It is of 6 digits..as Frank pointed out.. 20th to 25th column.

When I run with your data and use 20,6,PD, it works. When I run using 20,5,PD, it gets an 0C7 ABEND as expected because the sign is invalid in that case.

Quote: I have read in a lot of places that
Statement 1:
SORT FIELDS=COPY
SUM FIELDS=NONE will remove the duplicates.

I don't know where you read that, but it wasn't written by anyone who knew what they were talking about.

SUM FIELDS=NONE can remove duplicates with:

Code:
SORT FIELDS=(p,m,f,s,...) or
MERGE FIELDS=(p,m,f,s,...)

SUM FIELDS=NONE cannot remove duplicates with:

Code:
SORT FIELDS=COPY or
MERGE FIELDS=COPY or
OPTION COPY

SUM FIELDS=NONE is only valid for sorting or merging, not for copying. SUM FIELDS=NONE removes records with duplicate keys. Since copy does NOT specify a key, SUM FIELDS=NONE cannot be used to remove duplicates when copying. Sort and Merge specify a key, so SUM FIELDS=NONE can be used to remove duplicates when sorting or merging.

Quote:
Statement 2:
SORT FIELDS=NONE will also avoid duplicates.

Code:
SORT FIELDS=NONE

is invalid syntax. It will result in an error message.

lanand_hps · Joined: 05 Dec 2007 Posts: 59 Location: chennai

Thanks a lot Frank .!