DSXchange

Are your VarChar() columns bounded or unbounded? Using unbounded strings will save space, at a cost of performance.

Ahh, that is correct. the INDEX function won't work as a filter condition. Instead of a FILTER stage, use a transform stage with the INDEX() function in the constraints.

I just did a dummy test job with ExecSH of "echo #DSProjectName#" and that output my project name correctly. This was at V8.

My guess is that your file has the first line as headings, and you didn't specify that in your sequential file stage.

The Index function you gave will work. What is your probelm with it?

When I searched the forum for the keyword "39202" I found many threads relating to this error. Have you checked to see if any of the suggestions in those threads might help you?

Such global actions can often be done by editing a .dsx export file and using macros or other tools to make the necessary changes. Note that this could potentially destroy a project, so needs to be done carefully and with backup of the existing project contents.

The default is 50 Warnings, unless changed in the Director -> Options; I am not aware of a project level setting in the Administrator.

Would it be possible to remove the character conversion from the process and stick with numeric values? This would make it simpler (and faster).

This was asked yesterday, check this thread.

Buffering is why. I would guess the stage was "Finished" but the job was still running "Running" as well.

Embedded quote characters in strings need to be doubled in order to be correctly read in again, i.e. the string would need to look like in order to represent the string

We didn't want to be child-killers, we just throttled them a bit; sort of like Homer does to Bart

saikir - could you explain how the "Create File" can degrade performance, as I haven't observed that. Sometimes, with large hashed files, the CLEAR.FILE command can take a while; I've found that deleting and re-creating the file can be faster and also will take care of disk space used in the physica...

Inserting into hashed files can get slower as the file grows. Are you certain that you delete and re-create the hashed file each run?