Improve indexing for multi-lingual items

[ruthtillman]

Our language faceting is based on the primary language field 008[35-37]. The 041's subfields contain additional language data. We have not been indexing this data because it can be overkill, e.g. https://searchworks.stanford.edu/view/13749584 (The Crown showing English, Arabic, Danish, Dutch, Estonian, Finnish, French, German, Italian, Norwegian, Polish, Spanish, Swedish, Turkish languages.)

But if the 008[35-37] is mul for multilingual, the 041a will often be useful. The 041a may repeat. It contains the same language codes as the 008[35-37], so the mapping won't need to be updated. But this will turn the facet from a single value into an array/list.

Some records will not have an 041a, so we'll need to do an "if exists" check.

So new logic would be:

• If 008[35-37] == mul:
• THEN index the 041a fields too
• AND make unique (because the 041a may contain "mul" as well)

We'll want to retain the "mul" as well.