parallel processing

[vij]

Hi,

I have this scenerio : I want to use a Dataset as the target in my job, which gets loaded in the first stream/pipeline of my job.

In the same job, I want this target as a source and load another dataset.

Is it not possible to do it in a single job?
Becos I got an error and wen i spliited the job, the error was solved...


Also, can somebody explain the INDEX function with an example?particularly, wen it will return 0?

[aakashahuja]

Returns the starting position of a substring.

Syntax
Index (string, substring, instance)

string is the string or expression containing the substring. If string is a null value, 0 is returned.

substring is the substring to be found. If substring is an empty string, 1 is returned. If substring is a null value, 0 is returned.

instance specifies which instance of substring is to be located. If instance is not found, 0 is returned. If instance is a null value, it generates a run-time error.

Examples
The following examples show several different ways of finding the position of a substring within a string:

MyString = "P1234XXOO1299XX00P1"
Position = Index(MyString, 1, 2)
* The above returns the index of the second "1" character (10).
Position = Index(MyString, "XX", 2)
* The above returns the start index of the second "XX"
* substring (14).
Position = Index(MyString, "xx", 2)
* The above returns 0 since the substring "xx" does not occur.
Position = Index(MyString, "XX", 3)
* The above returns 0 since the third occurrence of
* substring "XX" * cannot be found.


P.S:- This is a BASIC function.

[ArndW]

Aakash - the INDEX function syntax is the same in PX, so it will work on both types of jobs.

[Nageshsunkoji]

Hi,

In Parallel it is not possible to use the same dataset as a Source and Target. You required to create a new job for this functionality. Where as incase of server we have HASH file facility to perform the same.

[ray.wurlod]

In the same job, I want this target as a source and load another dataset.

Is it not possible to do it in a single job?

No. "Blocking" operations such as this are specifically prohibited in parallel jobs, as they prevent pipeline parallelism from occurring.