DSXchange

The job dependency tab on the properties is a poorly thought out design by Ascential. In fact, I hear that the Sequencer is being overhauled again because it's sooo bad. Anyway, if you use a Sequence job to organize a job stream, it's supposed to document in each called job the dependency. What if y...

Why can't you simply walk each key recursively to the logical beginning of the linked list and just use the head of the list? No matter what, your issue is one of recursion, which is not handled well via a DataStage GUI metaphor job design. Your best bet is probably to do the recursive walk+join via...

Read the online documentation for full details. All links going into a link collector must share the same metadata.

Use the aggregator stage and choose the LAST method.

You need to be careful if you are doing a lot of repetitive calls to a DS function doing OPENSEQ statements looking for the same file. DS has issues and will false-positive hit sometimes. The most solid solution is to use the OS's own tools for file existence (dir on NT and ls or test on UNIX).

Now wait a second. If your job is seq --> xfm --> hash, then you should use a write-delay cache when loading this file. A write-delay cache means that while the data is loading into the hash file, it actually uses a memory mapped file and only once the file reaches the maximum cache size or the job ...

How wide is your average row? Did you begin with a minimum modulus? If so, you know it takes some time to initially create the hash file? Also, if the hash file is really wide? Let's get some numbers here to help us. How many rows, how many characters per row, how many columns. What is the file syst...

You could do: OCI --> seq --> hash, but, if you wanted to instantiate the OCI -->seq portion to achieve parallel sets of queries spooling results, then having this as an all-in-one job doesn't let you. Let's say you have a column in the query that you could do a MOD(columnname, 5) on. If you have 5 ...

Do it in job control before your run your job. Simply use a DOS batch script using the dir command and check for the results of the file. Use the DSExecute API (see your manual) and either call it using a custom BATCH job control, or use it in a DS user function/routine.

Break your job into two separate jobs. Job #1 spools from the source to a sequential file. Measure your rows/second. Job #2 reads the sequential file and loads into the hash. Measure your rows/second. Job #2 should astound and amaze you. You should see the speed at which DS can take a simple sequent...

No.

The auto-generated SQL produced by the OCI stage for 8i and 9i slaps a TO_CHAR with formatting of YYYY-MM-DD HH24:MI:SS on every column defined as DATE or TIMESTAMP. If you're not using the OCI stage or you're not on 8i+ or you're using USER-DEFINED SQL of some nature, than Oracle will return a date...

No. There is no way to communicate transformer stage variables with before/after stage routines. You do have a method using the @USER? and COMMON variables (search the forum).

DataStage has no concept of data types, everything, especially dates, are to be considered as string. That being said, your constraint comparison is lexicographic, and because your format is incompatible with that type of comparison, you do not get the result you desire. In DataStage, to compare dat...

Use stage variables (search the forum) to pivot your many rows into one. You'll have to add an extra row of data into your primary input stream to signal the end of the data to output the last row. Your data must be sorted and grouped. Another method is to write to a hash file, referencing the row b...