See the Instance (top) portion of this wiki page: https://folio-org.atlassian.net/wiki/display/FOLIJET/Classification+and+Call+number+handling+in+Data+Import

Overview: When an incoming MARC Bib record has multiples of the same field in some call number fields, the data in the Instance classification fields is not parsed correctly

Current workaround: Only one of each subfield must be in the call number fields of the incoming MARC Bib record

Steps to Reproduce:

1. Log into FOLIO-snapshot as diku_admin

2. Import the attached MARC file using the default Create Instance and MARC Bib job profile

3. Once imported, click on the completed file, and then click on the "Created" hotlink in the Instance column

4. Check the classification section of the first instance record (where multiple $a are in a single field)

Expected Results:

• When call numbers are in multiple separate fields (see the second record in the attached file), everything works fine

• See the column called Required handling when subfield in one field is repeated If $a is repeated, then the second $a (and any following $b) should be separated to a new classification field in the Instance

• This already works properly for call number types UDC (080) and Gov Doc (086)

Actual Results: 

• When call numbers are in multiple separate fields (see the second record in the attached file), everything is fine

• When multiple subfield a's are in the same field, default mapping does not work properly for call number types LC (050 and 090), NLM (060) and Dewey (082)
  There are all the values from the multiple 050 in the one field