I have recently done some "house-keeping" looking for consistency in the MySQL db (5.5.24 in 3.8.5), and while some tidying up has been straight-forward I am having problems with more_subfields_xml in items. Only about 90% of 'items' have any xml attached to them. Within those that do have an xml entry, only the following subfields appear: <subfield code="f"> entries < 1% <subfield code="i"> entries ~ 20% <subfield code="k"> entries ~ 1.5% <subfield code="x"> entries ~ 95% and they are mostly filled with random values from 2xx, 6xx and 9xx, although the 'code="x"' values appear to be valid from the 952$x. code="f" contains values from 952$v; code="i" has some 952$i valid entries but mostly 650$a,x,y,z; code="k" has random authors, titles, etc - and I can't find where we might use 952$k in our frameworks. Can someone please tell me how these blobs are mapped/created and used by koha? Our odd-ball results/entries don't appear to be having much, if any, influence on cataloguing and search functions. Also, if there is any way to "rebuild" more_subfields_xml -- it's way too big a problem to manage manually. Many thanks - Paul P.S. For completeness, after de-duping all the entries obtained from " SELECT more_subfields_xml FROM items; " (249,636 text lines), the only data left (apart from the subfields above) are "empty containers" using: xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns="http://www.loc.gov/MARC21/slim"> xsi:schemaLocation="http://www.loc.gov/MARC21/slim http://www.loc.gov/standards/marcxml/schema/MARC21slim.xsd" <?xml version="1.0" encoding="UTF-8"?> <collection </collection> <datafield tag="999" ind1=" " ind2=" "> </datafield> <leader> a </leader> <record> </record>