DSXchange

:? So... you are successfully converting a date string in YYYY-MM-DD format to an internal date using what you posted, yes? And now you need to insert the result into SQL Server. What is the datatype of your target database column - DATETIME, SMALLDATETIME, something else? What is the "SQL Serv...

About the only "precaution" I can think of would be to shut them down gracefully if you have the chance to do that.

What is the "command activity" doing? If it is only being used to determine what happened with Job2, remove it as you don't need it. The conditions you posted should be in the triggers coming from Job2.

Actually, can you post the exact syntax you are using rather than paraphrasing?

Where are you getting this "internal format" from and how exactly are you attempting to convert it? Post the syntax you are using.

They are being made "safe"... and you beat me to posting the answer. Dang. Yup, that tells it that it is already XML and to not futz with it any further.

Interesting... thanks for posting that little tidbit.

Have you read the documentation on that function? That number is the internal value of today's date. Use OConv() if you need a date string in a specific format.

You would need to create a custom routine to go get whatever error messages you wanted from the logs of the jobs the sequence ran. Up to you then how you "throw" it after that. If you don't need the details but just the high-level list of what failed, that would be in the log of the sequen...

Can't really tell - did you try recompiling "job a"?

OK, had me going there for a moment. That much was acknowledged but the question here is specifically around the PX equivalent of the Server Folder stage. From what I've seen and what Ernie has blogged about, that equivalent would typically be the External Source stage. Do you concur or has you gots...

Looks like they really need to identify duplicates rather than remove them.

So... there's a Parallel version of the Folder stage in the 8.x release?

Three "lines" means there are three keys on that hashed file and those are the key values for the record that failed. Did you run out of space, perhaps? Does the hashed file get cleared each run or is it persisted?

Interesting... not an issue I have ever had, perhaps I just got lucky all these years.

Well, as noted in the linked thread you could do three things:

1. Ignore them as they are just informational
2. Turn off trace if you no longer need it on in those jobs
3. Resize the phantom directory