DI/DS Improvements & Suggestions

[dnewton]

Something that just popped into my head... the ability to drag a dataflow from the Central repo window onto a Job or Workflow workspace, and have DI prompt you with the Checkout options at that time.

Rather than making you go the central repo, check out there, then back to your local repo window, find the object, and drag that into your workflow.

[nitrobrother]

What happened to the idea of moving Join Rank to the query instead of (or in addition to) the tables as mentioned here? How about at the EDF level?

[wdaehn]

Still planned. Like so man yother things.

[dnewton]

[eganjp]

I'm in the middle of moving three projects for two different clients to production so I'm experiencing all the joys of coordinating the different install/configuration points.

It sure would be nice to be able to use the Management Console for everything. Once the job server and Management Console have been installed and are up and running I would like to be able to use the Management Console for:

1. Creating the repository or even upgrading
2. Configuring the job server to talk to the repository
3. Importing an ATL file to the repository - no more designer required for production.
4. Selecting System Configuration settings.

The biggest issue with going to production is that the operations support people often don't know anything about Data Integrator. They have to use so many DI applications to get a production instance working that it is painful.

1. Install the job server and management console (Unix)
2. Install Designer (Windows)
3. Create database (Oracle)
4. Configure job server (Unix)
5. Import ATL file using Designer (Windows)
6. Change System Configuration using Designer (Windows) or change Datastore connections using Designer or the Management Console.
7. Export the job execution command for external scheduling (Unix or Windows)

[djordan]

It would be great if I could disable/enable only certain parts of a job when testing rather than having to create a separate 'test job'.

[Edefius]

Hello Everybody!

I don't know if Is it possible to create folders in Local Object Library??
I think no, I think It would be great if we can for example sort "data flow","work flow",... per project/subject

I would find this option very practice and my data service projects would be more easy to organise.

Edefius.

[dnewton]

To expand on Jim Egan's post above:

I would like to see all DI configuration and administration available within the Web Admin. This would include:

1. DSCONFIG
2. Job Server setup
3. Repo maintenance/setup
4. Calculating dependencies/mappings
5. Ability to stop/start DI services

Since we're asking for the moon, let's also have this rolled into the BO Enterprise CMC entirely, so that there's a single security model across all BO products... That way we can use AD authentication and not have to maintain separate account schemes. And then you can let the BOE installation process worry about web/app servers, Java dependencies, etc. Items 4 & 5 above would be easy to do. The other 3 would require a user interface, so I don't know how hard it would be to extend BOE to do that.

[dnewton]

Give a super-admin the ability to undo a checkout in the central repo... At the moment, you have to hack the database with SQL scripts, which is a time-waster and is dangerous.

[dnewton]

More options for setting the Degree of Parallelism either dynamically (based upon target partitions) or globally at the Job level at run-time/schedule-time:

http://www.forumtopics.com/busobj/viewtopic.php?t=125523

[adrian.rutter]

The ability to specify an Oracle Home to use for a DataSource would be useful.
If the server is Oracle 10g, but has the need to connect to an older version of Oracle Server if I install an Oracle 8.1.7 client it would be nice to have the ability to stipulate the oracle Home to pass the SQL Net statement down.

Old database connection - Oracle 8.1.7 client home
Other database connections - Default home.

[dnewton]

Two ideas about the Central Repo:

- Ability to do a "show where used" on an object in the central repo, and have it list other places in the central repo where used (not the local repo)

- Ability to tag objects in the central repo as "deprecated" (unused/retired), which would show them as greyed out. (And if you did a Show Where Used and that deprecated object was still being referenced, it would note it as such in the dialog box.)

[Darth Services]

The ability to display the attributes date created / data modified for an object in the Local Object Library (in the list of column headers) and sort by them so I can find out what has most recently been worked on or new additions without going through each items properties!

Further to this the ability to add additional attributes to objects. Such as category, and display/sort/filter by these in the Local Object Library (which also covers grouping mentioned by previous posters).

[dnewton]

Another minor UI/usability/productivity improvement:

In the Sys Config Editor… you can click on the Configuration Name column header, and sort by config name. That’s good. However, the other columns in the grid are sorted in no particular order.

If you have many data sources, and want to quickly find one source out of that many, it requires a lot of searching. The datasources in the grid should be presented sorted in alphabetical order, column-wise.

(ADAPT01199544)

[dnewton]

While using the Web-based Management Console, looking at the Batch Job status... I'd like to have a hover-over menu on the Job Name, that had options for:

- Show History - this would link you over to the Operational Job Execution History page (/odreports/odJobList.do) with that particular job already filtering the results

- Auto Documentation - this would link you over to the Auto Doc section (/smartdoc.jsp) with that job already chosen


It would save some clicks.