|
View previous topic :: View next topic
|
| Author |
Message |
anirudh1985
Currently Banned
New User
Joined: 22 Aug 2007
Posts: 66
Location: bangalore
|
|
|
|
Hi All,
I need to develop a tool in REXX. The desc is as below:
Generally when a field is expanded in a file, other fields have to be shifted towards right to accommodate the new field (expanded). This utility should expand the field and as well as reformat the file to accommodate the new field.
For Example : ID X(5),Name X(20),PH X(10),CITY X(10), COUNTRY(10)
If City field is expanded to X(20), the already existing data in the COUNTRY field has to be moved towards right to accommodate the 10 characters newly added.
Can anybody help on this as i am new to rexx
Thanks in advance! |
|
| Back to top |
|
 |
enrico-sorichetti
Superior Member
Joined: 14 Mar 2007
Posts: 10863
Location: italy
|
|
|
|
Here is a snippet...
| Code: |
/* inp contains the old data
ID X(5),Name X(20),PH X(10),CITY X(10), COUNTRY(10)
*/
id = substr(inp, 1, 5)
Name = substr(inp, 6,20)
Ph = substr(inp,26,10)
City = substr(inp,36,10)
Coutry = substr(inp,46,10)
/* reformat each field to the proper length */
City = left(City,20," ")
/* out contains the reformatted record
ID X(5),Name X(20),PH X(10),CITY X(20), COUNTRY(10)
*/
out = Id || Name || || Ph || City || Country |
|
|
| Back to top |
|
|
Pedro
Global Moderator
Joined: 01 Sep 2006
Posts: 2453
Location: Silicon Valley
|
|
|
|
| You also have to worry about the record lengths. The new record might not fit in the old file. Be prepared to reallocate the dataset to a new size. |
|
| Back to top |
|
|
enrico-sorichetti
Superior Member
Joined: 14 Mar 2007
Posts: 10863
Location: italy
|
|
|
|
but for huge amount of data I would investigate the dfsort/icetools capabiliies,
it will outperform rexx by a factor of five
it is not that difficult to move things around |
|
| Back to top |
|
|
anirudh1985
Currently Banned
New User
Joined: 22 Aug 2007
Posts: 66
Location: bangalore
|
|
|
|
thanx a lot!... |
|
| Back to top |
|
|
anirudh1985
Currently Banned
New User
Joined: 22 Aug 2007
Posts: 66
Location: bangalore
|
|
|
|
| enrico-sorichetti wrote: |
Here is a snippet...
| Code: |
/* inp contains the old data
ID X(5),Name X(20),PH X(10),CITY X(10), COUNTRY(10)
*/
id = substr(inp, 1, 5)
Name = substr(inp, 6,20)
Ph = substr(inp,26,10)
City = substr(inp,36,10)
Coutry = substr(inp,46,10)
/* reformat each field to the proper length */
City = left(City,20," ")
/* out contains the reformatted record
ID X(5),Name X(20),PH X(10),CITY X(20), COUNTRY(10)
*/
out = Id || Name || || Ph || City || Country |
|
How can this be generalised so that for any given record structure and for any given field it has to be expanded by desired length and reformatted? |
|
| Back to top |
|
|
enrico-sorichetti
Superior Member
Joined: 14 Mar 2007
Posts: 10863
Location: italy
|
|
|
|
If You' d explain what You are trying to achieve
You might get better answers
Your last requirement was a bit unclear...
are You asking about data reformatting ?
in this case as I already said rexx will not give top performance
and most of the transformations ( like the one explained before )
can be carried out by sort/icetool |
|
| Back to top |
|
|
anirudh1985
Currently Banned
New User
Joined: 22 Aug 2007
Posts: 66
Location: bangalore
|
|
|
|
[quote="enrico-sorichetti"]If You' d explain what You are trying to achieve
You might get better answers
I need to develop a tool which will take any file say VSAM and expand any of its filed to the specified length. The file can have 10 fields or 20 etc... it has to work for files with varied fields.. i hope i am clear.... |
|
| Back to top |
|
|
enrico-sorichetti
Superior Member
Joined: 14 Mar 2007
Posts: 10863
Location: italy
|
|
|
|
| in what format are You planning to give the description of the input and output layouts ? |
|
| Back to top |
|
|
anirudh1985
Currently Banned
New User
Joined: 22 Aug 2007
Posts: 66
Location: bangalore
|
|
|
|
| enrico-sorichetti wrote: |
| in what format are You planning to give the description of the input and output layouts ? |
my input will be a copybook with the record format of the file for which a field has to be expanded and a VSAM file whose fild has to be expanded and output has to be a reformatted VSAM file.... |
|
| Back to top |
|
|
enrico-sorichetti
Superior Member
Joined: 14 Mar 2007
Posts: 10863
Location: italy
|
|
|
|
Hi Anirudh!
two copybooks I would suggest...
the format of the copybook would be cobol I guess
also remember that REXX does not have any VSAM access facility
( unless You use external packages , not free )
so You must repro the vsam to a sequential file
carry on the reformatting
repro to the reformatted file
i am still of the opinion that rexx is not the best suited tool for the task
do You realize that You have to build a small cobol parser scanner to achieve the mapping You want,
You must decode all the possible variable layouts and definitions
...take my word for it, I do know about compiler and interpreter writing
that' not a simple task, especially if, no offense meant, You must rely on forum replies to implement it
I just looked at the IBM file manager specs and here is a quote of the
copy feature
| Quote: |
You can use the Copy Utility to:
* Copy data from any partitioned, sequential, or VSAM data set to any other partitioned, sequential, or VSAM data set.
* Select the records to be copied using the start key (VSAM only), skip and copy count fields, or a conditional expression defined in an input template.
* Change file attributes. You can copy records where the input and output data sets have differing record formats, record lengths or block sizes. The copy process truncates or pads records appropriately. To specify a pad character to be used, use the PAD processing option on the Set Processing Options (option 0) panel.
* Copy selected fields, change the size of fields, and create new fields in the output file by using a "From" template with a "To" template.
* Allocate a non-VSAM data set, or define a VSAM data set.
* Copy sequential data sets or PDS or PDSE members, converting the data to or from ISPF PACK data format. |
I guess that fileaid if present in Your shop should have the same capabilities
did You look at the rlcl sample I pointed You to |
|
| Back to top |
|
|
Garry Carroll
Senior Member
Joined: 08 May 2006
Posts: 1184
Location: Dublin, Ireland
|
|
|
|
| Quote: |
also remember that REXX does not have any VSAM access facility
( unless You use external packages , not free ) |
There is cbt268 which gives REXX access to VSAM.
| Quote: |
do You realize that You have to build a small cobol parser scanner to achieve the mapping You want,
You must decode all the possible variable layouts and definitions |
The cobol copybook parser (cobdfsym) in DFSORT's Sort Tricks document (sorttrck.pdf) would probably be useful for this.
I agree that REXX is probably not the best way to achieve the object. It may be that something clever using DFSORT with generated SYMNAMES for the 'before' and 'after' copybooks could be used to provide a solution.
Garry. |
|
| Back to top |
|
|
mbizzle
New User
Joined: 06 May 2020
Posts: 7
Location: United States
|
|
|
|
I registered just so I could share this fix to COBDFSYM.
I had been using it for a few years now, and loved it and was sad when our upgrade from cobol compiler V4 to V6 bricked this program. Luckily, it was a fairly simple fix to update it.
Here are the fixes you need to make to allow it to be compatible with the V6 (and i assume V5) compilers.
I am posting this on a few forums so that others may find it that have the same issue.
change this line
parse var line 92 asmdef datatyp .
to
parse var line 87 asmdef datatyp .
and add this line
do until left(line,1) = '1'
call Put_line
parse pull line
if pos('End of Data Division Map',line)<> 0 then leave
end |
|
| Back to top |
|
|
|
|
