DSXchange

Not ideally; VarChars on average are shorter than the maximum length specified. But if you total the lengths, you'll err on the side of caution, and over-size your hashed file. This is far better than under-sizing it.

Sleep? Wazzat?

It probably will make no difference under default conditions, as the two Lookup stages would probably be combined (they share the same parallelism, and are combinable). The advantage of two Lookup stages would be that you could separately identify the lookup failures from each source. Bear in mind t...

kumar_s wrote:

When I read the file in parallel

Are you referring this to multiple readers per node? If so its not recommended for variable length delimited file. Its for fixed width files.

I think that all that was meant was "in a parallel job".

Just place the command in the Command field of an Execute Command activity in the job sequence. The command output will contain the required number. You may need to trim field marks. If you want to parameterize the file name, put wc -l in the Command field and the job parameter reference (with surro...

dsdlockd does not kill processes (phantoms). It merely clears locks held by defunct or non-existent processes. Hence the "lock" as part of its name. The licensing tool ($DSHOME/bin/dslictool) can be used to recover licensing, and may allow defunct processes to die, but that's not guaranteed either.

It's Saturday in Australia.

Total all the characters in a row and add 1 for each column. There are no data types - numbers are stored as character strings.

Does your UNIX support +3 option for the head command and the +4 option for the tail command? If so, you have a very easy filter you could use to get the detail rows only.

"First line is column names" will ignore only one row, not three.

Do you have to stage into tables? Sequential files are much faster for staging areas. Are you REALLY still on version 4?

Use the Hashed File Calculator to tune the mkdbfile command. This utility is on your DataStage clients CD in the Utilities\Unsupported folder.

As far as I am aware most ODBC drivers for Oracle end up using the OCI (that is, the "full client"). Certainly this is the case with the non-wire Data Direct drivers that ship with DataStage.

Not necessarily, if the export included the job executables.

Do you have the Field Width property set on these truncated columns? What data type have you assigned to them?

That's why I asked you to report all 16 job start times. All 16 finish times would be useful also.