Tricky Requirement

[vinay_renu2001]

Hi

I have a requirement as follows:

We have a Directory called "REP" in which there are 200 Sub-Directories each for one customer. In each Sub-Directory they have 4-5 reports generated bi-weeekly for every customer by Oracle. In those reports few of them are generated in PDF format and few of them are generated in .TXT format. All these Reports need to be Archived at some different location, Need to convert all the .TXT Reports into PDF foramat and need to attach these Reports to an E-mail and forward them to corresponding Customers. All the Sub-Directories have the standard naming convention(e.g: mnc1100, mnc1210,mnc2555....). Reports does not have any standard Naming Convention.

Note: Currently All the above process is performing manually.

Now my requirement is to automate all the above process using Shell Script and DataStage. Here in this scenario, i dont think DataStage helps us a lot, Am i correct?

Iam ambigous to handle this task in a folder level or a file level.

Can anyone give me any suggestions!!!

Thanks in Advance...
Viny
"You must be the change you wish to see in the world."
--Mahatma Gandhi

[ray.wurlod]

You are correct. This is not something that DataStage does. But DataStage can control the shell script(s) that you create to perform the movement of files. You can use an Execute Command activity from a job sequence, or you can use a before/after subroutine from a job, for example ExecSH (the task of which is to execute a command in /bin/sh).

[kumar_s]

Yes, there is nothing much you can make use of Datastage other that Execute Command Activity from which you can call the unix levelconverter for these sort of house keeping of moving from one directory to another. I would not use import/Export operator for this. Unix level 'move' would be more prompt from the shell.

-Kumar

[djm]

If you end up having to write a UNIX shell script as part of your solution, one command that you are likely to find useful is "find" ("man find" for more details). Although this can be a tricky command to master, it is very powerful. For example

Code: Select all

find /REP \( -name "*.TXT" -exec ls \{\} \; \) -o \( -name "*.PDF" -exec ls -l \{\} \; \)

will look in all sub-directories underneath /REP and execute the command "ls" for a file with a suffix of .TXT and execute the command "ls -l" for a file with a suffix of .PDF.

Obviously the command in the example does not do anything near what you are looking for. However, I would suggest trying something simple commands until you work out the meaning of the various flags ( e.g. -exec, -name) and shell escape sequences ( e.g. \( \; ). You should then be able to tackle building a more complete solution (e.g. for files named with a ".TXT" suffix, invoke your own script to convert it to a PDF).

David

[ray.wurlod]

It could be done quite straightforwardly using DataStage BASIC. I refrained from answering in this vein earlier because the question is posted in the parallel forum.

The DataStage BASIC (possibly a custom job control routine) would execute a UNIX command (such as find or tree) to return a treewalk of the subdirectory structures, then generate the appropriate commands to archive the different file types and execute those commands.

But of course this could also be done with a shell script, probably just as easily. It really depends on one's skill set.