Hash&Sequential

[kamalraaj]

I am designing my jobs to read a file and get all the parameters value to start the job. Is it good idea to use Sequential file or Hash File to store values which we are going to use for all the jobs. If we use this file for all our jobs will there be any problem with the Share Lock when multiple jobs runs at the same time. Which is the effective way for this problem.

Thanks
Kamal

[ray.wurlod]

Both ways are effective if none of the parallel (simultaneous) jobs is going to update the file. Share locks do not apply when a file is opened for reading, unless there is a writing process. From your description, the writing process has finished at this point.
Sequential files have the advantage over hashed files that they are accessible (visible, readable) from the operating system.
Hashed files have the advantage over sequential files that they can be cached in memory. However, for a load of parameter values, the size in either case is unlikely to be more than one or two memory pages, so it's going to be resident in any case.


Ray Wurlod
Education and Consulting Services
ABN 57 092 448 518

[WoMaWil]

Remember Kamal, that you generaly have mostly 3 different kind of information:
- environment information which change with your environment (Development, Test, Integration and Production
- general parameters , which change very rarly
- daily changing parameters like actual Jobname, actual date and so on.

We have made good experiences with having divided the files in this three parts, because we won't need to touch things which never change.

Wolfgang