Group comparison/update between two files HeaderVsdetail

premnath · New User Joined: 08 Jan 2006 Posts: 15

File1:(VB/600 length)
Header1 ==> Identified by X'22003A' in position 19 ->Key info available here
Rec1 - No update
Rec2 ==> identified by X'153A27' in position 19 (This where final update should happen)
Rec3 ==> identified by X'153A27' in position 19 (This where final update should happen)
Rec4 ==>- No update
Trailer1 ==> dynamic position wherever we get X'220D0A'
Header2 ==> Identified by X'22003A' in position 19 ->Key info available here
Rec1 -- No update
Rec2 ==> identified by X'153A27' in position 19 (This where final update should happen)
Rec3 ==> identified by X'153A27' in position 19 (This where final update should happen)
Rec4 ==> identified by X'153A27' in position 19 (This where final update should happen)
Rec5- No update
Rec6 - No update
Trailer2 ==> dynamic position wherever we get X'220D0A'
Header3 ==> Identified by X'22003A' in position 19 ->Key info available here
Rec1- No update
Rec2 ==> identified by X'153A27' in position 19 (This where final update should happen)
Trailer3 ==> dynamic position wherever we get X'220D0A'

File 2:(VB/600)
Header1 ==> Identified by X'22003A' in position 19
Rec1: ==> identified by X'153A28' in position 19 -->Key info available here
Rec2
Trailer1 ==> dynamic position wherever we get X'220D0A'
Header2 ==> Identified by X'22003A' in position 19
Rec1:-
Rec2:==> identified by X'153A28' in position 19-->Key info available here
Trailer2 ==> dynamic position wherever we get X'220D0A'

Requirement: Compare File1(header info) and File 2(detail info) based on below criteria and update file1 with information that is
fetched from file 2
Comparison Positions :
File 1(Key info available in Header in below positions)
Pos 1 To Pos 4: eg:X'1537'(STORE NUM)
Pos_22 To Pos_23 : eg:X'0668'(Terminal Number)
Pos_30 To Pos_31 eg:X'0079'(Transac Num)
Pos 28 To Pos 39 X'TT0000000001' (token)
Pos 43 X'59'(Indicator y/n)
File 2(Key info available in detail record in below positions)
Pos 24 To Pos 25 X'1537'(STORE NUM)
Pos 27 To Pos 28 X'0668'(Terminal Number)
Pos 30 To Pos 31 X'0079'(Transac Num)
Pos 37 To Pos 48 X'TT0000000001' (token)
Pos 50 To Pos 61 X'PP0000000001'(This is not key to compare ,this is data for updation in file 1 when a match found)

Condition to be coded

IF <Pos 1 To Pos 4> OF fILE 1 header record = <Pos 24 To Pos 25> of file 2 Detail record
AND <Pos_22 To Pos_23> OF fILE 1 header record = <Pos 27 To Pos 28> of file 2 Detail record
AND <Pos_30 To Pos_31> OF fILE 1 header record = <Pos 30 To Pos 31> of file 2 Detail record
AND <Pos_28 To Pos_39> OF fILE 1 header record = <Pos 37 To Pos 48> of file 2 Detail record
AND Pos_43 OF fILE 1 header record = X'59'(INDICATOR 'Y')
THEN
Move <Pos 50 To Pos 61> of Matched detail record(From File2) TO <Pos 28 To Pos 39 OF DETAIL record>(To File1)
MOVE indicator value 'N' (hex) in Pos 43 of File 1(Detail record)
END-IF

Note :
Header of file1 Identified by X'22003A' in position 19
Detail of file1 Identified by X'153A27' in position 19
Header of file2 Identified by X'22003A' in position 19
Detail of file2 Identified by X'153A28' in position 19

I tried writing a cobol program for this ,But could not get full expected results as i was able to update only one record in group
if multiple details record has to be updated in a group i was unable to .Looking if there is a better way in dfsort/icetool.

Nic Clouston · Posted: Fri Oct 23, 2020 3:15 am

What have you tried? Please supply input/ouTput examples that actually look like your input and desired output - using the code tags to preserve spacing. Please use 'dataset' rather than 'file' as data is stored in datasets, not in files.

daveporcelan · Posted: Fri Oct 23, 2020 5:22 pm

Nic,

I was just thinking about this for some reason, and again you request the use of 'dataset' rather than 'file'.

I realize you are a global moderator with 2419 posts, but consider the possibility that you are incorrect on this point.

For this solution the user would use DFSORT, but in my example I must refer to the Syncsort manual as that is what I have handy.

Here is a snippet from the manual. Notice the reference to the word FILE several times. I searched the manual and the word FILE came up 532 times. The word dataset came up ZERO times.

What does that tell you?

The designer of the product refers to the collection of data as a FILE.

Why can't you?

Use the JOINKEYS statement to enable join feature processing and to identify the fields used to select records for join processing.
The join feature joins records from two input files that are specified on the SORTJNF1 and SORTJNF2 DD statements. By default, when the JOINKEYS fields from m records in SORTJNF1 match the JOINKEYS fields from n records in SORTJNF2, all combinations of the records are joined using the REFORMAT statement, producing m*n records as input to subsequent MFX processing. (This is called an “inner join.”)

The input files do not need to be presorted or have the same record type.
Two JOINKEYS control statements are required – one for each of the two files used in the join.

sergeyken · Posted: Fri Oct 23, 2020 5:58 pm

The question from NicC remains the same: what have you tried?

You typed tons of words, mainly meaningless, but did not present your own attempt to approach any solution, even a minor one.

daveporcelan · Posted: Fri Oct 23, 2020 6:31 pm

Rohit Umarjikar · Posted: Fri Oct 23, 2020 9:27 pm

Let us just focus on the problem than getting into files va Data set because I remember we hav had discussed this so many times in past and outcome is many would prefer both and some will prefer ONLY Dataset , so let us leave that choice to an individual what to use.
ibmmainframes.com/about66687.html

Besides , TS after 14 years in the forum don’t use code tags nor show us sample input and expected output data In neat way doesn’t get to any help so easily.

daveporcelan · Posted: Fri Oct 23, 2020 10:11 pm

Rohit Umarjikar · Posted: Fri Oct 23, 2020 10:16 pm

Technically its still a Data Set on zOS but I don't disagree if someone make a use of Files instead since IBM Manuals too refer many a time a file and Data set interchangeably.

Joerg.Findeisen · Posted: Fri Oct 23, 2020 11:37 pm

Besides the discussion to call datasets also files, I would like to see some input data here to play around with.

premnath · New User Joined: 08 Jan 2006 Posts: 15

Hi ,
Kindly ignore my previous request and consider this which is similar .By mistake posted it in different topic .Not sure about deleting that one
Need some help here in Group comparison .Your help will be much appreciated .Below is the details
Input file1 Type:VB
Input file2 Type:VB
Output file1 type :VB
Output file2 type :VB
All files Length:600
Input file1 Header identifier :X'22003A' in Position 19
Input file1 Trailer identifier: X'220D0A' it could be at any position after any number of records .Whenever
we encounter X'220D0A' that is identified as trailer record for those group of records .
Input file2 Header identifier :X'22003A' in Position 19
Input file2 Trailer identifier: X'220D0A' it could be at any position after any number of records .Whenever
we encounter X'220D0A' that is identified as trailer record for those group of records .
Input file2 Detail record Identifier : X'17' in Position 20
Requirement:Compare certain key fields from Header Record of Input file1 with that of
Any Detail record of File2 (that contains X'17' in Position 20)
and write all matching and unmatching entire group of records from Inputfile1
into o/p file 1 and o/p file 2 respectively
KEY Criteria for Comparison:
IF Pos 1 To Pos 4(Input file1 Header Record) = Pos 1 To Pos 4(Input file2 Detail Record)
AND Pos_22 To Pos_23(Input file1 Header Record) = Pos_22 To Pos_23(Input file2 Detail Record)
AND Pos_22 To Pos_23(Input file1 Header Record) = Pos_29 To Pos_30(Input file2 Detail Record)
THEN
Write only MATCHING group of records from Input file1 In output file 1
ELSE
Write only UNMATCHING group of records from Input file1 In output file 2
END-IF

Input file 1 sample:
Code:

Nic Clouston · Posted: Tue Nov 24, 2020 4:01 pm

Why could you not start a new topic as it is a new question? Now a moderator has to clean up the mess.