Search found 3045 matches

by vmcburney
Sat Aug 12, 2006 4:06 pm
Forum: IBM<sup>®</sup> DataStage Enterprise Edition (Formerly Parallel Extender/PX)
Topic: Migration from Abinitio to Datastage
Replies: 13
Views: 10403

I wouldn't be surprised if there was a single tool that had a conversion engine and had various adaptors for Informatica, DataStage server jobs and Ab Initio. In theory they could build a converter for any ETL tool if the demand was there.
by vmcburney
Sun Aug 06, 2006 10:35 pm
Forum: General
Topic: Necessary Evil?
Replies: 13
Views: 8447

kcbland wrote:Proctologists :shock:


I would call that a necessary discomfort rather than a necessary evil! They do have the patients best interest at heart (unlike the parking inspector).
by vmcburney
Sun Aug 06, 2006 5:56 pm
Forum: IBM<sup>®</sup> DataStage Enterprise Edition (Formerly Parallel Extender/PX)
Topic: dropping null columns in tranformer
Replies: 4
Views: 1177

Sounds like a good candidate for a custom built stage. Something that could navigate through the list of columns looking for nulls in a loop with an output link and a reject link. It is code intensive but you can do it in a transformer stage. Most functions will reject a row if that row has nulls in...
by vmcburney
Sun Aug 06, 2006 5:52 pm
Forum: General
Topic: Necessary Evil?
Replies: 13
Views: 8447

I have the definitive "necessary evil": parking inspectors.
by vmcburney
Sun Aug 06, 2006 5:49 pm
Forum: Site/Forum
Topic: Ascential DeveloperNet sunsetting this month
Replies: 20
Views: 10678

There are certain minor useability things that I think make or break a forum. For the old developernet the registration and login restrictions broke it and restricted it to 1/100th the volume of dsxchange. Getting a login was hard, once you had a login it did not automatically log you in and this wa...
by vmcburney
Thu Aug 03, 2006 2:25 am
Forum: IBM<sup>®</sup> DataStage Enterprise Edition (Formerly Parallel Extender/PX)
Topic: DB2 EE stage - rows getting rejected
Replies: 11
Views: 2490

Hey heah, I had forgotten about that blog entry! It even has a link to my wiki entry on retrieving SQL codes for DB2 stages! The other article of interest is Data Quality Firewall 2: inside an ETL job which has some bitmaps of how to pass your reject flow into a shared container.
by vmcburney
Thu Aug 03, 2006 2:23 am
Forum: IBM<sup>®</sup> DataStage Enterprise Edition (Formerly Parallel Extender/PX)
Topic: Advantages of PX comparing with other ETL tools
Replies: 3
Views: 1411

Ab Initio was the first of the parallel ETL tools and on scalable performance it is hard to beat. The main difference is that DataStage parallel is now more widely used, has easier access to training and documentation and online help and better integration with related data integration products. Inf...
by vmcburney
Wed Aug 02, 2006 9:35 pm
Forum: Site/Forum
Topic: Ascential DeveloperNet sunsetting this month
Replies: 20
Views: 10678

Yes, the developerworks forums are terrible. The interface is really, really bad. The search is poor and doesn't seem to work at all well and has no advanced search option. These are mirrored forums from real newsgroups so I assume there are newsgroup tools out there that we could use instead?
by vmcburney
Wed Aug 02, 2006 9:23 pm
Forum: IBM<sup>®</sup> DataStage Enterprise Edition (Formerly Parallel Extender/PX)
Topic: DB2 EE stage - rows getting rejected
Replies: 11
Views: 2490

There is a trick to it. Turn on colum propogation out of the DB2 stage reject link and type in the two column names manually. They will appear! For a peek stage it automatically shows all propogated columns such as the existing rejected columns and the two new database columns. Changing this to a se...
by vmcburney
Wed Aug 02, 2006 12:48 am
Forum: IBM<sup>®</sup> DataStage Enterprise Edition (Formerly Parallel Extender/PX)
Topic: Date Range lookup
Replies: 17
Views: 5948

The transformer stage has no reason to repartition data, it has no requirements for grouping key fields for example. The easiest design is to leave the transformer as default - same partitioning so it retains whatever the input link has set partitioning to be. If it is the first partitioning stage t...
by vmcburney
Wed Aug 02, 2006 12:30 am
Forum: Site/Forum
Topic: Ascential DeveloperNet sunsetting this month
Replies: 20
Views: 10678

On Tuesday I blogged about Developernet being unavailable and Developerworks not being ready yet. I actually got a blog comment from a US IBM employee when the forums were ready! It appeared at that point as though they had put it in the DB2 forums (which would have resulted in a very critical blog)...
by vmcburney
Tue Aug 01, 2006 5:06 am
Forum: IBM<sup>®</sup> DataStage Enterprise Edition (Formerly Parallel Extender/PX)
Topic: SCD In Parallel Jobs
Replies: 9
Views: 2905

There are two ways to get and apply your change capture. You start with a Before set of data and an After set of data. If you use the change capture stage alone it gives you new/changed/deleted/unchanged records. There is a flag to turn each one of these on or off and an extra field is written out t...
by vmcburney
Mon Jul 31, 2006 9:28 pm
Forum: IBM<sup>®</sup> DataStage Enterprise Edition (Formerly Parallel Extender/PX)
Topic: Date Range lookup
Replies: 17
Views: 5948

If I had DataStage running I would run a quick test to see. I am assuming that transforms are run first followed by constraints. You could test it by creating a transformer with a reject link, a constraint that blocked all rows and a transform function that triggered a rejection. If the constraint r...
by vmcburney
Mon Jul 31, 2006 7:11 pm
Forum: IBM<sup>®</sup> DataStage Enterprise Edition (Formerly Parallel Extender/PX)
Topic: Date Range lookup
Replies: 17
Views: 5948

The transformer is a good allrounder: offering constraints, transforms, stage variables etc. The problem is that it performs all those transforms on all rows before the constraints are applied. This means if your input to the lookup has 100mill records, your range lookup expands it to 300mill record...
by vmcburney
Mon Jul 31, 2006 3:56 pm
Forum: IBM<sup>®</sup> DataStage Enterprise Edition (Formerly Parallel Extender/PX)
Topic: Date Range lookup
Replies: 17
Views: 5948

That is spot on, you need a lookup stage followed by a filter stage. Resist the urge to put in a transformer with a constraint as it will be slower in this instance. Better to have a lookup - filter - transformer rather than a lookup - transformer. DataStage 8 (Hawk) adds range lookup functionality ...