DSXchange

Hi Guys,

How to create a job to load the data in a pre-populated Teradata table using Fast Load. Please give me guideliness for doing this job.As i am new to teradata with datastage.

Thanks in advance to helping out.

created

Is this a suggestion or do you really have any problem.

Teradata Enterprise Stage supports Fast Load in Append mode. Read the pertinent chapter in the Parallel Job Developer's Guide

ray.wurlod wrote:Teradata Enterprise Stage supports Fast Load in Append mode. Read the pertinent chapter in the Parallel Job Developer's Guide ...

Thank You.

You cannot fastload into an already populated table. it Multi loads data into the table. this is one of the reasons that a temporary table is created when we use the enterprise stage.

Teradata Certified Master V2R5

this is what the teradata enterprise stage does internally in the first place..................

Fast load cant load data into the allready populated table. To laod data into prepopulated table use the multiload( Best for the block level operation) or TPUMP (Provids real time facility)

where did i say it can ?

hamzaqk wrote:You cannot fastload into an already populated table. it Multi loads data into the table. this is one of the reasons that a temporary table is created when we use the enterprise stage.

What you say is true, but DataStage hides that from the developer by describing the operation as "Fast Load in Append Mode".

Thanks, but i do not support the idea that it should be called "Fast load in append mode" in the first place as the name conflicts with the real functionality of fastload itself. Apart from this the only fast load which the job does while using Enterprise Edition is into the empty table work_orch_xxx from which it simply does and select-insert opertation. (of record another very fast way of loading data in TD provided PI is the same for temp table and target table). :D

one (possibly useless) remark on using enterprise stage for TeraData - we found that it seems to do a count(*) before and after the actual load (which as already mentioned does a fastload into a staging table - then insert from subselect).

If there are more than 2Billion records in the target table then the count causes an aritmetic overflow which generates an error (though doesnt abort as I recall) - u cant switch this off - but can cause be confusing

I think the confusion here is around semantcis of the words "Fast Load" ; In the Teradata world, Fast load is the name of their utility that loads only to a non-empty table. When IBM developed the Teradata Enterprise stage, they worked with Teradata to make a fast load that can append to a table. So the DS Enterprise stage makes use of Teradata Fast Load to load a temporary (empty) table and then insert into the target table thus providing what looks like an Append.

Hope this clears up the confusion. So when you speak to the Teradata centric folks ,don't tell them Fast Load to append. That will drive them

Have a look at this blog post Datastage 7x Enterprise Editionwith Teradata , it has details about Teradata Utilities (& how) which can be used in DataStage.

Another reason it should not be called Fast load is because Fastload does not support loading into a table with secondary index on it.

Teradata Certified Master V2R5