DSXchange

Prashant,

The Px function BitXOR() will do this for you. If you do an XOR function of your test number with a number that has only the 17th (from the left) bit set to TRUE then if the result is TRUE that bit is a 0 otherwise it is a 1.

Kumar,

IBM/Ascential uses InstallShield for installation on the clients. You can download information on that product's error codes and other information from the InstallShield website.

Kishan,

the rollback segments are used by all Oracle processes; so it might be that another job was using a lot of space and then yours came along and demanded more than was available and thus got the error.

Installshield -113 means not enough disk space.

Nivas, increasing the APT_CONFIG from 4 to 6 doesn't meant that now you have 6 processes in total where you only had 4 before; it means that each stage in your job now executes with 6 instead of 4 jobs. So if you have 20 stages in your job this APT_CONFIG change has increased the number of process r...

Sumit, loading a UniVerse table is never faster than loading a hashed file. A UniVerse table is nothing but a view on a hashed file. Loading to a dynamic hashed file gets slower when writing large volumes of data because of misconfiguration, not because it is inherently slower. Splitting a hashed fi...

Nivas,

these parameters have nothing to do with the APT_CONFIG file; they are system parameters.

The two parameters that affect you in HP-UX are maxuprc (processes per user) and nproc (total processes on system). You cannot change these, they must be modified by a superuser on your system.

At unix, enter the command "db2 connect to <db> user <userid> using <password>" replacing the instance, user-id and password with the values you are using in your job and see if the connect works correctly. It might be that these settings have changed since the point in time when your job ran correc...

Your operator or System Administrator will need to make configuration changes to allow more concurrent processes available per user id. Each version of UNIX has a different way of doing this, so without knowing what platform you are running on we cannot help directly.

Can you connect from the UNIX shell using db2 connect and the same parameters you are using in the job.

Without NLS you can only use a 1-character separator and record delimiter. In your case the best solution depends upon how many records are processed and how many columns there are. One way would be to read the record as a single column, with no quotes and no column delimiters and using &H07 as ...

A shared container is DataStage's way of doing what a compiled language does in subroutines. It does not make sense to make a shared container, or even a normal container, for a source read stage unless the exact same query with the same columns and parameters is used everywhere. The documentation i...

Your input file type is of PL/1 and you can import the metadata into the Manager and use it to read the file. This is not a complex file, so it is easy to load.

Sekr,

my first thought is that if you don't want the data, don't ask for it. Remove the extra columns from your query so that they are not processed. But perhaps I have mistaken your intent on your 2 posts.

Cherry, a question phrased that way sounds so much like an interview question that those who have used both product will probably not be interested in responding. Why not state why you are looking for answers to this question and be satisfied with just one reason? The answer is so judgemental as wel...