DSXchange

Either your userid doesn't have permissions to open the file in the mode you need, or the DataStage jobs point to disk locations where there is no file available. You need to research that yourself, perhaps going into the designer and comparing programmed with actual file locations (you can use view...

Andal - I've been trying to tell you several times now that these two attributes cannot with 100% accuracy determine the datatype of a column! If the justification is L instead of R it is probably text versus a number. If the conversion is MD<n> then it is most likely a number since the MD conversio...

HTTPS is not a file transfer protocol, it is a secure HTTP protocol layer. The two really are different beasts. There are secure FTP and secure copy mechanisms available for both Windows and Unix that are better suited for this kind of job.

You need to have a sequence or some other controlling agency to trigger this, since as you've noticed it cannot be done from within the job itself.

From my post earlier - you can't tell directly if a column is char or integer from the hashed file definition. You should work on the column's contents instead of its definition since hashed file contents are actually unstructured.

On AIX you need to have Visual Age C++ 6.0, XL C/C++ v7 installed.

Hashed files have different data types to SQL. A hashed file column has two optional attributes to specify data representation: 1. The format/justification such as "12L" or "4R", and 2. An attribute for a conversion such as "D4/E" or "MD2". The combination of these two attributes gives the effective...

Andal, I am confused by your requirements. A hashed file is usually read using a hashed file stage, not with an SQL query (mainly for performance reasons). The column type is not an attribute in either a table or a hashed file stage that can be read. It is possible to read the hashed file's DDL by u...

Greg,

that date is awfully close to the -4713 minimum date supported by Oracle timestamps. Does a date of year -4711 show up correctly?

This easy to do using 2 stage variables in a transform stage. Declare
StageSequence as IF In.ID=LastID THEN StageSequence+1 ELSE 1
LastID as In.ID

If you do the same export and import as a .dsx file does it work without an error? If yes, then you need to report this bug to IBM/Ascential through your support provider. If both methods show the same error then it bears further investigation, you can use your favorite editor to locate that string ...

This sounds like a problem that needs to be submitted through your support provider - it looks like a bug.

Vinodand,

you have posted to the PX forum and this is a server issue
If you use indentation or at least autoformat your jobs in the manager it would make the jobs easier to read, maintain and this error most likely would have been visible right away.

Abby,

the error message ds_seqopen() - error in 'open()' - Permission denied quite clearly means that the process does not have sufficient permission (either r or rw). Remember that the user process running the DataStage job might be different from your user.

If they are unrelated the performance difference between 3-in-1 or 3 separate jobs is negligible, they will process using the same number of processes at the same time, just a miniscule amount of overhead less since only one log file is being used.