SQLScript in SAP BW (AMDP Routines)

A transformation already has one AMDP (SQLScript) end routine. You now add a start routine — can you write it in ABAP?

With an InfoSource between two transformations, which mixed ABAP/HANA arrangement is allowed?

How many AMDP classes get generated for a transformation's routines — BW on HANA vs BW/4HANA?

A BW transformation contains an ABAP start routine. Your team wants it to run in the SAP HANA runtime for speed. Can it?

How does a DTP running in the SAP HANA runtime move data differently from one running in the ABAP runtime?

Default DTP package size: ABAP runtime vs SAP HANA runtime?

In BW/4HANA you set a debugger breakpoint in the transformation's _M AMDP class. It never triggers. Why?

How does an AMDP transformation routine's code reach the target system, given the generated class can't be transported?

What are the four parameters every AMDP transformation routine method receives?

In a BW/4HANA 2.0 routine you reference ERRORTAB, but activation fails saying it's unknown. What's missing?

In a BW AMDP transformation, what is the generated source-record key field for, and what must you do with it when a routine creates new rows?

Your end routine does no error processing, so you assign only OUTTAB and leave ERRORTAB alone. Result?

You join a /BIC/ table in an AMDP routine. Where do you write the name with double quotes, and where without?

To look up data from another aDSO in a routine, should you read /BIC/A<name>2 (the active table) directly?

An end routine LEFT OUTER JOINs a lookup table and writes to an aDSO. Non-matching rows fail at load. Why, and the fix?

You read a custom (non-BW) client-dependent table in an AMDP routine and get rows from every client. What went wrong?

Start routine vs end routine — what fields does each see in INTAB/OUTTAB, and when does each run?

Field routine B needs the value field routine A computes. Why can't B just read A's field, and what do you use instead?

You create an expert routine and try to activate the transformation from SAP GUI. It fails with syntax errors. Is the tool broken?

Porting an ABAP field routine (called per record) to AMDP, you write it to loop over each row. What's the AMDP mental model instead?

Error handling via the error stack in a HANA-runtime transformation isn't working. What's the version-sensitive reason?

With error handling on, does a routine's runtime double because the DTP runs it twice?

You set the sort field like this. Records come out ordered 1, 10, 11, 2, 3… Why? ``` RANK() OVER (ORDER BY d) AS "RECORD" ```