Accented characters not appearing correctly in BI

问题:

Using Oracle Business Intelligence Publisher 12.2.1.3.0 AKA, BI Publisher Enterprise.

I'm having trouble with accented characters, such as the ü character (the letter U with an umlaut/diaeresis). Sometimes they appear with the umlaut above the u, and sometimes to the right of the u.

I've learned that a u with an umlaut/diaeresis can be stored as

One character (composed)

latin small letter u with diaeresis

Two characters (decomposed)

latin small letter u
combining diaeresis

I originally thought it was a problem with the html2fo function, but that has proven incorrect. Some of my source data is stored composed, some decomposed.

The problem is with the decomposed form.

It's working fine when I generate HTML output, but appears incorrect in PDF output. PDF output shows the umlaut after the U, rather than above it.

PDF Output:

If I take the source data in Word, and export a PDF, the decomposed characters appear just fine.

How do I correct for this problem? Bug in BI Publisher maybe? I would not like to make a function to convert decomposed to composed characters if I can avoid it.