Updating the input file

sanjayis01 · New User Joined: 13 Jun 2007 Posts: 55 Location: banaglore

I want to update the input file

Requirement :

Input file is VSAM file. Get the particular column from the VSAM file and copy in to a sequential dataset.

VSAM file DATA

ABCDEFGHIJKLMNOPQRSTUVWXYZ

USING OUTREC FIELDS GET THE PQR IN TO FLAT FILE .

FLAT FILE DATA

PQR

NOW I WILL EDIT THE FLAT FILE DATA.. (SAY SOME OTHER APPLICATION WILL MODIFY THE DATA)

NOW FLAT FILE WILL HAVE

INSTEAD OF PQR IT HAS 123 ..

NEXT I WILL USE THIS FLAT FILE HAVING DATA 123 I WILL BE USING INPUT IN THE NEXT STEP AND UPDATE THE POSTION IN THE VSAM FILE HAVING DATA SHOULD BE REPLACED BY

EXISITING: ABCDEFGHIJKLMNOPQRSTUVWXYZ

THE FINAL OUTCOME SHOULD BE

:ABCDEFGHIJKLMNO123STUVWXYZ

PLEASE LET ME KNOW IF YOU NEED MORE DETAILS

expat · Posted: Tue May 20, 2008 11:33 am

Why go to the bother of unloading a subset of the file, updating that subset, reconstructing records to put back into the VSAM file and then actually returning them to the VSAM file.

Why not just process the whole record using SORT or even a program to update in place.

1) Backup of file
2) Update of file

But ...............
Will every record be updated ?
Will the record key be part of the unloaded subset ? If not, how will you know where each updated subset goes in relation to the VSAM key ?

sanjayis01 · New User Joined: 13 Jun 2007 Posts: 55 Location: banaglore

The Things is the system which updates tha will support the flat file not vsam file and the and reclength should be 3000 .. the vsam file is having length more than 3000.

thats the reason getting the data to flat file and making 3000 length to call the opther system to update the fields..

so i would say this is the best approach .. using program we can do it.. but

i want it by JCL .. this will increase the produtvity

enrico-sorichetti · Posted: Tue May 20, 2008 5:51 pm

let' s see if I got Your requirement right

You have a dataset ( it' s organization is irrelevant for the logic )

whenever at a certain offset/position for a certain lenght there is a certain string,
You want to substitute that string with something else...
right ??

in other words from Your sample records

sanjayis01 · New User Joined: 13 Jun 2007 Posts: 55 Location: banaglore

i Have a file

file 1

1234567890123456789

I will get the field starting @ position 10 - 14 of length 5 in to seperate file FILE 2 using OUTREC FIELDS

now File 2 will have ( here i am editing it manually ,, but some other apllication will use this file and update the fields)

01234

Now i will edit the File 2 with

ABCDE

So the updated field in the file 2 has to replace the postion mentioned in the FILE 1 i. e 10 - 14

so o/p file should look as

123456789ABCDE56789

Please let me know how i can achieve this ..

please let me know if you need more details

sanjayis01 · New User Joined: 13 Jun 2007 Posts: 55 Location: banaglore

Please have look @ this

expat · Posted: Tue May 20, 2008 6:57 pm

Enrico,

This is the bit that gives me some serious concerns.....

OP can only process records of 3000 bytes, but the VSAM records are longer than 3000. I think that this will require a programmed solution to merge the download amended 3000 byte records with the correct 3000+ part of the record still in the VSAM file.

sanjayis01 · New User Joined: 13 Jun 2007 Posts: 55 Location: banaglore

That i can manage i will conver vsam to flat file ..

please consider this as simple flat file ..

please let me know the solution...

enrico-sorichetti · Posted: Tue May 20, 2008 7:04 pm

You did not reply to my question !!
for the change process do You already know the from string and the to string

or that info is hidden in some other database/dataset

sanjayis01 · New User Joined: 13 Jun 2007 Posts: 55 Location: banaglore

Could you please clear doubt again ???

i did not get ..

i think i have given all the information . i think what u have understood previously its correct..

enrico-sorichetti · Posted: Tue May 20, 2008 7:18 pm

the difference is between having to run four steps instead of two

if You know the from string an the to string

You can run a
dfsort step to create a dataset with the changed columns/fields
dfsort step to reload the vsam

in the other case You have to run
dfsort step to extract the from string
some_user_program to build the from to relation strings
dfsort step to create a dataset with the changed columns/fields
dfsort step to reload the vsam

sanjayis01 · New User Joined: 13 Jun 2007 Posts: 55 Location: banaglore

please let me know how i can achieve this

dfsort step to create a dataset with the changed columns/fields

sanjayis01 · New User Joined: 13 Jun 2007 Posts: 55 Location: banaglore

Hi All

Please let me knw the solution

enrico-sorichetti · Posted: Tue May 20, 2008 9:48 pm

I do not want /cannot give a working sample because I cannot test it right now

if You search the dfsort/icetool forum You will find lots of samples
here is a pointer to a thread that looks good
http://ibmmainframes.com/viewtopic.php?t=30503&highlight=ifthen+overlay+copy

snippet taken from there not tested
adjust the offset to Your need and exercise on a small dataset first

sanjayis01 · New User Joined: 13 Jun 2007 Posts: 55 Location: banaglore

The problem is in the above example u mentioned is

They know the what data they are getting in the input file.. i dont know what data is coming and by what i wil be updating ..

every time i will be getting different number only . i only know the psotion

so in that case how i can handle it..

Skolusu · Posted: Tue May 20, 2008 10:23 pm

sanjayis01,

You can create a DFSORT symbol to update the field. However you need to retain the old value also to be able update it once again. So when picking the the bytes to be replaced we create 2 files

sanjayis01 · New User Joined: 13 Jun 2007 Posts: 55 Location: banaglore

Hi

thanks for reply , i think this could work for me . i will test it and post the reply ..

mean while i got few question .. please let me know the answers for the below

could you please explain this again

//ACTUAL DD DSN=YOUR.ACTUAL.VALUE FILE,DISP=(,CATLG),SPACE=(TRK,(1,1),RLSE)
//UPDATE DD DSN=YOUR.UPDATE.VALUE FILE,DISP=(,CATLG),SPACE=(TRK,(1,1),RLSE)
//SYSIN DD *
SORT FIELDS=COPY
OUTFIL FNAMES=ACTUAL,
BUILD=(C'ACTUAL_FIELD,C',C'''',10,5,C'''',80:X)
OUTFIL FNAMES=UPDATE,
BUILD=(C'UPDATE_FIELD,C',C'''',10,5,C'''',80:X)

For what purpose this actual file is used..?

and also please explain me this card

BUILD=(C'ACTUAL_FIELD,C',C'''',10,5,C'''',80:X)

how can i mention the psotion in this ?

Skolusu · Posted: Tue May 20, 2008 10:53 pm

The actual field is used to update the input file where you picked the value as you are overlaying the contents with something else

In the above control cards I am picking the value at pos 10 for 5 bytes

sanjayis01 · New User Joined: 13 Jun 2007 Posts: 55 Location: banaglore

OUTFIL FNAMES=ACTUAL,
BUILD=(C'ACTUAL_FIELD,C',C'''',10,5,C'''',80:X)
OUTFIL FNAMES=UPDATE,
BUILD=(C'UPDATE_FIELD,C',C'''',10,5,C'''',80:X)

i am not familiar with this syntax please explain me in brief what each fiels indicates in the BUILD CARD.

Skolusu · Posted: Tue May 20, 2008 11:37 pm

C'ACTUAL_FIELD,C' will populate a constant ACTUAL_FIELD,C at pos 1

C'''' will put a single quote following the last field

10,5, will copy the contents from pos 10 for a length of 5 bytes

C'''' will again put another single quote

80:X will pad up to 80th byte with spaces

the output will look like this

sanjayis01 · New User Joined: 13 Jun 2007 Posts: 55 Location: banaglore

Thanks lot.. man

Its working nicely...

hats off..

sanjayis01 · New User Joined: 13 Jun 2007 Posts: 55 Location: banaglore

Suppose i want update multiple column in the same way means

how can i proceed????

Skolusu · Posted: Thu May 22, 2008 9:05 pm

sanjayis01

From the sysout you have sent via private message , it appears that you are using syncsort.

I'm a DFSORT developer. DFSORT and Syncsort are competitive products.
I'm happy to answer questions on DFSORT and DFSORT's ICETOOL, but I
don't answer questions on Syncsort.