IBM Mainframe Forum

by **mahesh100** » Mon Aug 31, 2009 6:15 pm

Hi,

I need to remove the padded spaces from one of the field.

File layout
two-char x(2)
three-char (3)
name x(35)
indicator 9(1)
indicator 9(1)

Input file

Select all

HEADER
IN,IND,INDIA ,1,0
GB,GBP,GREAT BRITAIN POUND ,1,O
TRAILOR

(Remove all the extra spaces in the name field)
Output file:

Select all

HEADER
IN,IND,INDIA,1,0
GB,GBP,GREAT BRITIAN POUND,1,0
TRAILOR

I was thinking with Program but if can be done using DFSORT then it will be great.
Please let me know of any thoughts.

by **Frank Yaeger** » Mon Aug 31, 2009 9:38 pm

You description of the layout doesn't match your example. Your layout describes fixed fields, but your example has comma delimited fields. Which is it?

Also, what is the RECFM and LRECL of the input file?

Is the name field the only one that can have embedded blanks?

by **mahesh100** » Tue Sep 01, 2009 8:54 am

Frank,

Input file :
File layout
two-char x(2)
Filler value comma
three-char (3)
filler value comma
name x(40)
filler value comma
indicator 9(1)
filler value comma
indicator 9(1)

1) The file is delimited by comma.
2) Input file is fixed length and LRECL=51
3) Yes only Name field will have embedded spaces.

Thanks,
Mahe

by **dick scherrer** » Tue Sep 01, 2009 10:13 am

Hello,

Do not post the same question in multiple parts of the forum. . .
ibm-cobol/topic2319.html#p9079

But still i am thinking on it ,how to use that ...
If any help on that will be great...

What is not clear? The input is a "not quite usable" delimited string (because of embedded spaces). If the string is unstrung into several variables, the meaningful length of each component can be calculated (field length minus trailing spaces), and then the values can be strung back together using reference modification, eliminating the unwanted trailing spaces.

by **mahesh100** » Tue Sep 01, 2009 7:13 pm

Dick,

I wanted to use the ICETOOL or sort instead of program.
I am not sure why they want using sort, I guess it performance issue.

Thanks,
Mahesh

by **Frank Yaeger** » Tue Sep 01, 2009 10:35 pm

Here's a D[size=150][/size]FSORT job that will do what you asked for:

Select all

//S1 EXEC PGM=SORT
//SYSOUT DD SYSOUT=*
//SORTIN DD DSN=... input file (FB/51)
//SORTOUT DD DSN=... output file (FB/51)
//SYSIN DD *
OPTION COPY
INREC IFOUTLEN=51,
IFTHEN=(WHEN=(3,1,CH,EQ,C','),
BUILD=(1,7,8:8,40,JFY=(SHIFT=LEFT,LEAD=C'"',TRAIL=C'"',LENGTH=42),
50:48,4),HIT=NEXT),
IFTHEN=(WHEN=(3,1,CH,EQ,C','),
OVERLAY=(8:8,46,SQZ=(SHIFT=LEFT,PAIR=QUOTE)),HIT=NEXT),
IFTHEN=(WHEN=(3,1,CH,EQ,C','),
FINDREP=(IN=C'"',OUT=C''))
/*

by **mahesh100** » Wed Sep 02, 2009 9:01 am

Thanks Frank.
I will try this code.

by **mahesh100** » Wed Sep 02, 2009 8:47 pm

Thanks Frank... Its working Fine

But i did not get the logic of below line. Could you please let me know on this(What is the purpose of giving 50:48,4)

BUILD=(1,7,8:8,40,JFY=(SHIFT=LEFT,LEAD=C'"',TRAIL=C'"',LENGTH=42),50:48,4),HIT=NEXT),

Thanks,
Mahesh

by **Frank Yaeger** » Thu Sep 03, 2009 12:43 am

We are adding quotes around the name which increases its length by 2. We need to shift the indicators after the name before we squeeze them together.

For example:

Select all

GB,GBP,GREAT BRITAIN POUND ,1,O

would be justified to:

Select all

GB,GBP,"GREAT BRITAIN POUND" ,1,O

then squeezed to:

Select all

GB,GBP,"GREAT BRITAIN POUND",1,O

Finally, the quotes are removed and we get:

Select all

GB,GBP,GREAT BRITAIN POUND,1,O