IBM Mainframe Forum

by **lostcontinent** » Thu Jan 20, 2011 1:30 pm

hello again,

I have a big file with database content, so that variables containing binary fields are displayed in a "readable" hex-form:

Select all

01E4
VAR1 = '069B'
VAR1234 = 'G'
XYZ = '0674'
*PADDING = ' '
0E
VAR2345 = '37'

Is it possible to manipulate this output with sort to get the 2 lines containing the hex-code in 1 line in decimal?
In the above example:
VAR1 is '0016E94B'h, so it should be printed out as 1501515 (decimal)
VAR2345 is '03E7'h, so it should be printed out as 999 (decimal)

OR

Is it possible to get/keep the the content of the previous sort line?

frank

by **Frank Yaeger** » Thu Jan 20, 2011 11:39 pm

I'm not sure you covered all the possible variations for your input records, but assuming that you did, here's a DFSORT job that will extract the two line variables and display them in the way I think you asked for. Note that you'll need z/OS DFSORT V1R10 PTF UK90025 or z/OS DFSORT V1R12 PTF UK90026 (Oct, 2010) to use TRAN=HEX. If you don't have this PTF installed, ask your System Programmer to install it.

Select all

//S1 EXEC PGM=SORT
//SYSOUT DD SYSOUT=*
//SORTIN DD *
01E4
VAR1 = '069B'
VAR1234 = 'G'
XYZ = '0674'
*PADDING = ' '
0E
VAR2345 = '37'
//SORTOUT DD SYSOUT=*
//SYSIN DD *
OPTION COPY
INREC IFTHEN=(WHEN=GROUP,BEGIN=(10,1,CH,EQ,C' '),RECORDS=2,
PUSH=(21:14,4))
OUTFIL INCLUDE=(1,3,CH,EQ,C'VAR',AND,21,1,CH,NE,C' '),
IFOUTLEN=25,
IFTHEN=(WHEN=INIT,
BUILD=(1,11,51:21,1,14,1,22,1,15,1,23,1,16,1,24,1,17,1)),
IFTHEN=(WHEN=(55,1,CH,EQ,C' '),
OVERLAY=(61:51,4,TRAN=UNHEX,12:61,2,BI,EDIT=(IIIIIIIIIT))),
IFTHEN=(WHEN=NONE,
OVERLAY=(61:51,8,TRAN=UNHEX,12:61,4,BI,EDIT=(IIIIIIIIIT)))
/*

SORTOUT would have:

Select all

VAR1 = 1501515
VAR2345 = 999

by **lostcontinent** » Fri Jan 21, 2011 3:58 pm

Frank,
thank you very much. wo do not have the PTF installed (and it will take some time until it's done),
so I got another solution:
In a first step, I did some grouping and splicing, so I got a temporary file containing only the needed records with
the "2-record-hex-values" spliced together in one record with a special sign '$', so I can identify the records:

Select all

FUNK = 'GETD'
PRNAME = 'PVTG1A1 '
ANZQ = 1
VTN = $ 01EC0698
Ausdruck der gelesenen Projektionszeile:
VTN = $ 01EC0698
PSB4300 = '02'
PSB4040 = '20011201'
TEST = $ 01 A6'
EDAT = '20011130'
EZEIT = '13373574'
PSB4004 = $ 01760808
PSB4041 = '20011231'
PSB4042 = 'N'
>>>>> OK <<<<<
FUNK = 'GETN'
PRNAME = 'PVTG1A1 '
ANZQ = 1
VTN = $ 01EC0698
>>>>> NFND <<<<<

Now in a second step I converted the hex-codes (indentified by the '$' sign) to decimal:

Select all

//SYSIN DD *
SORT FIELDS=COPY
* deal with 4 digit hex number
INREC IFTHEN=(WHEN=(17,1,CH,EQ,C'$',AND,21,2,CH,NE,C' '),
OVERLAY=(31:C'0',19,1,
C'0',23,1,
C'0',20,1,
C'0',24,1,
C'0',21,1,
C'0',25,1,
C'0',22,1,
C'0',26,1),HIT=NEXT),
* deal with 2 digit hex number
IFTHEN=(WHEN=(17,1,CH,EQ,C'$',AND,21,2,CH,EQ,C' '),
OVERLAY=(31:C'000000000',19,1,
C'0',23,1,
C'0',20,1,
C'0',24,1),HIT=NEXT),
* convert hex digits in decimal values
IFTHEN=(WHEN=(17,1,CH,EQ,C'$'),
FINDREP=(STARTPOS=31,ENDPOS=46,
INOUT=(C'0A',C'10',C'0B',C'11',C'0C',C'12',
C'0D',C'13',C'0E',C'14',C'0F',C'15')),HIT=NEXT),
IFTHEN=(WHEN=ANY,
BUILD=(1,16,C' ',19:(31,2,ZD,MUL,+268435456),ADD,
(33,2,ZD,MUL,+16777216),ADD,
(35,2,ZD,MUL,+1048576),ADD,
(37,2,ZD,MUL,+65536),ADD,
(39,2,ZD,MUL,+4096),ADD,
(41,2,ZD,MUL,+256),ADD,
(43,2,ZD,MUL,+16),ADD,
(45,2,ZD),TO=ZDF,LENGTH=10))
/*

That works Ok for me:

Select all

FUNK = 'GETD'
PRNAME = 'PVTG1A1 '
ANZQ = 1
VTN = 0001501640
Ausdruck der gelesenen Projektionszeile:
VTN = 0001501640
PSB4300 = '02'
PSB4040 = '20011201'
TEST = 0000002582
EDAT = '20011130'
EZEIT = '13373574'
PSB4004 = 0001601640
PSB4041 = '20011231'
PSB4042 = 'N'
>>>>> OK <<<<
FUNK = 'GETN'
PRNAME = 'PVTG1A1 '
ANZQ = 1
VTN = 0001501640
>>>>> NFND <<<<

Next question: How can I get rid of the leading zeros?
If I put an EDIT=(IIIIIIIIIT) in the BUILD-command it gives me a syntax error?

thanks a lot in advance,
frank

by **lostcontinent** » Fri Jan 21, 2011 5:40 pm

solved:
I put an additional '$' directly after the decimal value and then:

Select all

* get rid of leading zeros
OUTREC IFTHEN=(WHEN=(17,1,CH,EQ,C'$'),
BUILD=(1,18,19,11,JFY=(SHIFT=LEFT,PREBLANK=C'0')),HIT=NEXT),
* check if JFY gave just a $, then replace it with 0
IFTHEN=(WHEN=(17,1,CH,EQ,C'$',AND,19,1,CH,EQ,C'$'),
BUILD=(1,18,C'0'),HIT=NEXT),
* replace $ with blank
IFTHEN=(WHEN=(17,1,CH,EQ,C'$'),
FINDREP=(INOUT=(C'$',C' '),STARTPOS=17,ENDPOS=30))

if there is a better/easier way to do this, please let me know ;-)

IBM Mainframe Forum

combine 2 lines in one with dfsort?

combine 2 lines in one with dfsort?

Re: combine 2 lines in one with dfsort?

Re: combine 2 lines in one with dfsort?

Re: combine 2 lines in one with dfsort?