DSXchange

Hello Craig, no, that doesn't address the issue. What I am writing is utility to analyze locks on this system and to get rid of spurious locks caused by network issues terminating sessions which the deadlock daemon doesn't catch. But in order to do that I need to trace back froma Device/inode value ...

The PORT.STATUS command outputs a DEVICE and INODE. I know that there is some method of XORing or doing some similar manipulation to that number in order to get a major and minor device number but cannot recall how to do that and have played around with reverse-engineering the displayed device numbe...

What about using the lineage and impact analysis functionality built into the product?

The PX canvas works differently than the Server canvas, a stage that writes to a sequential file cannot be simultaneously be used to read that file.

You cannot split a Dataset with a single command. There are two ways that I can think of:
a) Write a job to read the 1 dataset and create multiple ones
b) Use a script and the orchadmin subcommands "copy" and "truncate" or "dump" to prune and copy the data.

I am not sure how the database stage would get the same parameters in different nodes, as those parameters will be values from the data and thus each instance would be separate and distinct.

DataStage will run the procedure sequentially, per processing node. So if you had a 25-way configuration and used round-robin partitioning you would get parallelism. Otherwise option (1) would be the way to go. Both assume that the stored procedure is more time-consuming than the DataStage overhead ...

DataStage is backwards-compatible, but not necessarily forwards-compatible. You can do an export in V8, then edit the .dsx file so that the header information shows the one required for V7 (I'm not at a DataStage installation, so can't give you the exact numbers); the import at V7 might show warning...

Kim - that's pretty much what I remembered as well. A BASIC program that gets all top-level entries from DS_JOBOBJECTS and adds the appropriate entries to DS_JOBS should be sufficient.

kaps wrote:Can anyone answer this question of whether we need to have same version mentioned in the docs...

It must be the exact version mentioned.

No, RowProcCompareWithPreviousValue is a server-based SDK routine and will not work in PX.

nagarjuna - remember that $DSHOME is not necessarily set in most environments; that is why I didn't specify that. The more reliable method is to use something like "cd `cat /.dshome`"; also, I think that the dsenv might not need to be sourced in order to execute dsrun

That is a UNIX error, the path to the "dsjob" command is not in your environment's PATH. Either add it, or change your current working directory to that directory.

As Craig has suggested, put in a Transform stage that outputs a row down a link which has a constraint < 1 when "IsNull(In.ColA) AND IsNull(In.ColB) AND IsNull(ColC)"

Craig - you are correct, I didn't read the complete post and thought that the original poster just wanted column names. As you correctly noted, the derivation is not available as part of the link information; and thet DSGetStageInfo() routine call does not have an option to return derivations. So on...