DSXchange

Sreenivasulu wrote:U can use the REPLACE function of Datastage to do this.

I was really reacting to this post, of course!

Welcome aboard! :D Best practice is never to abort, so that you retain control. Pre-process the data to look for violations. If any is found, your job sequence can choose not to run the "real" job. Or to run an intermediate job to correct those violations, if such action is appopriate/possible.

@NULL and @NULL.STR are DataStage BASIC system variables and therefore inappropriate in parallel jobs. What kind of "input stage" are you using? For example, the Reject link from a Sequential File stage only handles rows that do not match the schema associated with the input link. You need to follow...

From memory, the data browser flips to exponential when the number of decimal places exceeds 4, for float and dfloat data types. This is certainly the case with certain calculations in server edition. Within DataStage BASIC this behaviour can be overridden via a PRECISION declaration, but we don't h...

from within Designer, press F1. In the search box, type in BASIC. One of the returned objects will be "Alphabetical listing of BASIC commands". There you will find vast amounts of information on any/all of the BASIC commands available in DS. Only some of them are in the on-line Help, ostensibly the...

Is the "timeout waiting to start" reliably reproducible?
What other load is on the machine at the time?

Nah! REPLACE replaces the value of an element in a dynamic array. You need to use CONVERT, EREPLACE or CHANGE. CHANGE is a subset of EREPLACE functionality. For single character replacement, CONVERT is the more efficient. However if you need portability to Windows platforms you will need EREPLACE or...

The Sort column contains 1, 2, 3 and so on to indicate the order in which the columns are known to be sorted; this would correspond to the order of column references in an ORDER BY clause. The Sort Order column has a drop down list from which you can choose Ascending or Descending as the order of so...

Source code for the before/after subroutine ExecDOS is supplied; you can inspect it and use it as a basis to create your own.

When you call the DSExecute() function the first, shell, argument, must be "DOS" on a Windows platform in order to execute a DOS-level command or BAT file.

... except B-tree indexes (usually)

They're probably stored OK (as float or dfloat?) but the viewer has a limited range of precision, outside of which it resorts to exponential view. That is, I believe that the restriction is in the viewer, not in the Data Set itself.

What did you find when you searched the forum for this?

Developers normally get granted CONNECT privilege. When it doesn't happen it's a gremlin. I have no other explanation available, but suspect some kind of timing issue.

Perhaps asking at ADN would prove fruitful.

That suggests that every row in the sequential file satisfies your condition, or that you have mistyped the constraint expression. Can you check the latter? Use the Debugger to step through a row at a time, inspecting (watching) the values of the attribute(s) as you step through. Set a breakpoint to...

You will get there, maybe with our help. People here have implemented these kind of things, for example I have implemented a solution with an arbitrary number of levels of dependency. It also incorporated a scheduler, and was all written in DataStage BASIC.