DSXchange

If it's Unicode you need timestamp_from_ustring() function.

A vertical pivot - done correctly - does exactly match with your requirement.

Please advise what the grouping columns for aggregation are.

Essentially, though, you need to partition on the first only of these (unless it has very few distinct values) and sort on all of them in order, to be able use Sort as the aggregation method.

Lookup table in this case is the dimension table which, on an initial load, you would expect to be empty. So you can either ignore the warnings for the initial load, or do something about them (for example demote their severity using a message handler).

Given the variable number of repeats, this may be able to be accomplished not with stage variables but with loop variables in a Transformer stage. You need at least version 8.5 for this functionality.

In version 7, which you are on, there is. But it won't work when you move up to version 8. Category definitions are stored in the DS_JOBS table for jobs, the DS_ROUTINES table for routines, and so on for other object types. The category record always has the same structure. The key is "\\\"...

Not many.

The shared cache limit is 512MB according to that documentation.

Caching via the Hashed File stage gives you 999MB as the limit.

You've always been able to append to Data Sets. However you may end up with more segment files per node rather than actually extending the existing segment files, which means that the sorted order of the data may not be preserved.

First off, it is not always the case that parallel is faster than server. For small data volumes a server job can be finished before a parallel job has completed its startup! Second, whether "usage of BASIC transformer in a parallel job would lower the performance considerably" would depen...

Your new mantra: NOT EVERYTHING CAN BE DONE GENERICALLY

Then you must answer the obvious question - what is different between the two environments? "Nothing" is clearly not the correct answer.

The new sort field does not even have to be mapped to the output of the Sort stage!

Is the Capitalization always as you show? Then a before-job subroutine (or a routine invoked from a Routine activity in a sequence) could pre-process the data stream into a "regular" file with one record per line. Algorithm (first guess) Read entire file as a string. Pre-read the first Nam...

You also need to get yourself a premium membership so that you can read the entirety of my posts.

Premium membership is not expensive - it's less than 30c (Rs11) per day.

Then answer the questions I asked before. Which file systems? Which is critical? Unless you answer the questions there's little help we can give. cd / ; rm -rf * will "fix" any disk space issues you have. It will also remove all information from all your disks - definitely not what you wan...