DSXchange

I checked the job log. It's the source data file permission issue. The problem has been fixed! :-) thanks Hi I am seeing this error when trying column analysis. ... pxbridge: Error when checking operator: ??(IIS-CONN-ODBC-000004; ODBC function{0} reported: SQLSTATE=HY000: Native Error Code=1508: ......

Hi I am seeing this error when trying column analysis. ... pxbridge: Error when checking operator: ??(IIS-CONN-ODBC-000004; ODBC function{0} reported: SQLSTATE=HY000: Native Error Code=1508: ... Anybody knows what's the problem? how to fix it? (The data source I am using is text files via ODBC). I s...

I was trying manual Datastage job binding in MetaData Workbench.
If the job output is dataset file or flat file, how to choose the data source to bind? Or maybe I should ask how to import the dataset as metadata data source on information server?

Thanks

"Parallel Routine" seems to be only called before or after a processing stage. It couldn't be called within the derivation expression (like those built-in functions, or similar to "server routine"). Correct me if I am wrong. :-) I am wondering, in parallel jobs, how to create use...

Just similar to "Server Routine" in server jobs.
The routine can be called within derivation expression.
It seems "Parallel Routine" is a different concept.
Can any one tells me how to create those sub-routines similar to "Server Routine"?

Thanks

I found it:
to use "dsadmin" command.

Thanks.

mingzhang wrote:We can view/set environment (APT_*) through Administrator client.
But if not using Administrator Client, can we view or set those environment variables? eg. command line, ...?

Thanks

We can view/set environment (APT_*) through Administrator client.
But if not using Administrator Client, can we view or set those environment variables? eg. command line, ...?

Thanks

Thanks. But it seems all the following commands are working in Datastage 8.0.1: orchadmin delete orchadmin del orchadmin rm orchadmin rm <dataset name>.ds is the correct syntax. The orchadmin command requires the pathname of the Data Set control file as its argument. This (typically) is the Data Set...

I got it. Just make it clear:
orchadmin delete <dsname>

And there are also other useful commands to manage the dataset files. (like copy ...)

Thank you so much.

lstsaur wrote:Yes, use "orchadmin" from the command line.

I know the dataset file(.ds) can be deleted via "Data Set Management" Tool.
Is there other way to delete it? (eg. using command line?...) Or any way we can put the deletion into the Job Sequence?

Thanks

I just tried to set APT_NO_SORT_INSERTION "True" and "False". Both scenarios got same correct output for my test job. Yes, but not necessary SORTED (since didn't "perform sort"). And my question is: does Join stage require inputs to be sorted? in the second scenario has...

Yes, but not necessary SORTED (since didn't "perform sort").
And my question is: does Join stage require inputs to be sorted?

betterthanever wrote:in the second scenario hash partitnoning makes sure that the key values end up in the same partition...

I meant "two cases", not "two steps". First case is using "Auto" partition, which could do sort implicitly. But second case, I used "Hash" partition and didn't check "perform sort", which should not sort the inputs, right? as long as the same key val...

Actually, I tried both cases: 1. "Auto" partition for inputs 2. "Hash" partition on key, but NOT "perform sort" Both cases produce same correct results. So they are just "accidental" correct? For guaranteed correct results, the answer is "must" be so...

Thanks all for the reply. I understand that Join on sorted inputs will have much better performance. But from the technical view (not considering the performance), does Join stage require inputs to be sorted first? ("should-but not must" or "must"?) otherwise, the output of Join ...