DSXchange

jreddy wrote:I tried DIRECT=TRUE for this job and it apparently is not committing anything until it reads the 92 mil records from source

There is no concept of intermediate commits in a direct path load, that is only applicable to a 'conventional' load because it does normal inserts.

Or generate a fixed name output file and then rename it after job. For example, we have a common script that date/timestamps a passed in filename and call it after job.

I don't see the need to see any code. You've got a Sequence job with a Routine Activity stage that returns a non-zero value and when 'automatically handle' is enabled without the proper triggers that is considered 'unhandled' and the job aborts. All documented in the online help and discussed here a...

Right. I meant define 'complete'. Session not active? Cursors closed? What does your DBA say is going on? Or is that you? Seeing as how you have OEM.

Mine understand. Need more redo/undo/transaction space? Just ask. Common stage that include all the possible columns involved? No. And just plain old stages with whatever columns you need, nothing special. Actually it's not clear how, having more (server) jbos with one OraOci per job can be maintei...

Sure, all kinds of things. Sort then use stage variables, use a hashed file, use an aggregator. Not to mention 'select distinct' from a source query. You'd need to get more specific for more specific answers.

Loading the sql at runtime from a file doesn't fundamentally change how queries behave. Meaning it will behave the same as if you'd entered that query in the stage itself, so that in and of itself isn't the issue here. I'm able to see the oracle session has completed ...but the control is not back t...

In the parameter's value, yes, as I doubt it needs to be a relative path.

dstage2006 wrote:and Scriptfile countfiledir/countfile as a parameters.

Scriptfile #countfiledir#/#countfile#

Unless the "scriptfile" is a parameter also, then:

#Scriptfile# #countfiledir#/#countfile#

You have no idea and the pushd/popd mask things. Why not comment out the "@echo off" for now and see what it logs?

Wow. My DOS memories are cautiously raising their heads from the depths I banished them to years ago. Start, pushd, popd... expanded memory, extended memory... ah, those were the days.

Nice. The important thing to realize is no constraint will fulfill the requirements and even adding an aggregator for a 'last' trigger would not guarantee 'only on successful creation'.

Then take that date and set the day to '01' for the 'first day of previous month date'.

jguerrero wrote:Do you try to use:

Command:#seq_path#\#APORIProcessDeleteFile#
Parameter: (what you need inside the script)

You can't because no parameter substitution is done in the Command box.

Did you ever talk to support? 7.5.2 on HP Itanium didn't install correctly and we had to create a symbolic link to get Oracle connectivity, hence the question.