DSXchange

... and away we go, split out for your dining and dancing pleasure. Consider yourself started.

We'll see if this gains any traction before deciding if it needs to be stickied or not.

LOL

If it is a duplicate value then it doesn't really matter which file it comes in from, does it? Unless, perhaps, you are tracking the source filename as part of the metadata. If not then a simple sort / remove duplicates could be used. The previous response assumes there will always be exactly three ...

Why not reverse the polarity? Meaning stream in the 10K and use the 1 record input file as the source for a hashed file lookup?

I think you've used up your allotment of exclamation points for the rest of the month.

william.wang_GMA wrote:Didn't notice those 2 buttons at bottom. I only saw that "Reply with Quote".

Not to worry, you are far from alone on that issue!

So, are we Resolved here? If so, please check the top of the page and mark the topic as such. Can help narrow down search results.

Just a quick FYI...

Please use the Reply to topic at the bottom (and top) of each page to reply. Leave the dreaded 'Reply with quote' for those few times when it is needed. Saves me having to come along behind people and scrape out all of the over-zealous over-quoting.

Good to know.

You are correct - it does suggest that and Ernie is known for writing large, expansive and comprehensive replies. In this case, however, that was all you got. :wink: Biggest thing was to clarify if you were discussing the client or the server software. Been out of the loop for awhile, so a question ...

You saw the entirety of his reply.

Amazon's S3 is a REST based service.

I for one don't require more details as I'm perfectly well aware of how your technique could / would be used, although I'm sure William will appreciate the write-up. I just don't think it applies here. At all.

OK, now I am officially lost. Ray has been answering your question about updating a parameterset via the associated values file but I don't believe that is germane to the issue at hand. Once the job starts, parameter values are 'locked in' and they cannot be changed. Typically, a job is deployed as ...

Yah... figured if you could have asked the original author, you would have done so already. Was really just trying to say in a polite way that I think you are SOL.

A "non-sparse" lookup will (typically) return the entire contents of the source table once in order to build a lookup cache locally on the DataStage server. The stage then executes individual queries against the local cache constrained by the Key columns passed in on your data stream. A &q...