DSXchange

What's the character set (code page) that the database is using?

You need exactly the same record schema that was used to populate the Data Set (from upstream of the Compress stage). In particular because that record schema includes tagged subrecords, you must have the fields that contain the tag values. These, according to the error message, are missing.

Turns out they've use functions rather than system variables. LastRecord() and LastRecordInGroup() are the names of the functions if I recall correctly.

Welcome aboard. Does the final field include newline characters?

Write out the logic in English. This will help you to understand what you need to implement using your query. It is better to learn understanding than to be handed solutions.

... and does it exist? Do you (the executing user ID) have appropriate permission to it and its parent directory?

You will need to use Lookup stage (or, possibly, Merge stage with appropriate unmatched masters handling). What don't you like about Lookup stage?

It seems apparent from the message that DB2 UDB API stage does not support reference links. Why not use a DB2 Connector stage?

This is a vertical pivot.

Techniques have been discussed here ad nauseam.

What is the data type of the lookup key field(s) on both inputs? Are strings identically trimmed? Cased?

Prove that it is "not executing".

Running on one node means that you're not tying up the process with waiting for database contention (locks).

Based on that admission you probably should get cogent help. You need to configure $ODBCINI and uvoodbc.config files, the latter in $DSHOME and in the project if you have one there.

The Surrogate Key Generator stage can not do anything except generate unique keys (provided you use the same source consistently) unless you re-initialize it. It doesn't matter what these keys are used for.

Use pairs of stage variables to "remember" the value from the previous row. Use the Sort stage to detect the beginning of a new group (a new sort key).

Are there earlier errors in the log, particularly from SQLAllocEnv() or SQLConnect() ? Can other jobs connect to this data source?