DSXchange

They shouldn't be expecting any formatting. Turn off the formatted output option you have on right now.

Four year old post.

Hopefully you've involved support in this, yes?

Since this is specifically about Sequence jobs, pushed this over to the General forum. Carry on.

Question: Does this behavior occur for every developer? Is this something that changed as in was fine but then went sideways or is this an always been this way kind of thing?

It's not unable to match it, it should be perfectly capable of doing that but you're right the CHAR part does make it a little trickier. I would suggest a couple of things, one or both may help. Test for the character value "161" rather than the number...i.e. use quotes so it does a string...

But if any application uses this sequencer it will generate positive values this is a drawback I was assuming this would be a dedicated sequence, something only used by the process in question. That's the only thing that makes sense to me. As to the attaching meaning part, I agree but I mentioned a...

At a basic level we're good with that definition. At a higher level there's the added concept of the data still being usable (appropriate) for the column and data type but that is de-identified / generic.

Waiting anxiously to see how this is all going to work.

Trust me, *I* know what masking is - just wanted to make sure that your definition matched the industry standard.

If by "mask the data" you mean you have to do transformations on data in specific (i.e. named) columns, then as Ray noted you can't use RCP for this effort. Perhaps an explanation of what you mean by "mask" would help.

These are the two I mentioned:

$APT_ORAUPSERT_COMMIT_TIME_INTERVAL
$APT_ORAUPSERT_COMMIT_ROW_INTERVAL

With those not set or not applicable and Record Count = 0 seemingly not working, I would involve support.

I don't think we are but if we're talking about an Oracle bulk load then there are two kinds, DIRECT and CONVENTIONAL. The former is the true bulk load where you have DIRECT=TRUE in the control file and there are no commits done as everything is written directly to the table and bypasses the redo me...

It has changed (well, been clarified) with every reply so far.

Best to start everything off proper and describe your full requirements in your initial post. That let's us get to the heart of the matter right away rather than all this back and forth, teasing out requirements one by one.

I was thinking a "count distinct" but the requirements just got opened up a little bit more than what was originally posted. My answer still stands but it seems more like a "double fork join" if there is such a concept, counts on both sides of the aisle that get taken and then co...

Generate it 'like normal' and then multiply it by -1.

Sure, related but it will change the Array Size to be an even multiple of the Record Count if it needs to but not the other way 'round. And that doesn't apply if the Record Count is set to zero. And as noted in the docs, if record count is a zero then "the connector processes all records before...