diff --git a/whelktool/scripts/dataimports/shb/.gitignore b/whelktool/scripts/dataimports/shb/.gitignore
new file mode 100644
index 0000000000..d5f83af524
--- /dev/null
+++ b/whelktool/scripts/dataimports/shb/.gitignore
@@ -0,0 +1,3 @@
+/build
+/cache
+/logs
diff --git a/whelktool/scripts/dataimports/shb/ConvertSHB.java b/whelktool/scripts/dataimports/shb/ConvertSHB.java
new file mode 100644
index 0000000000..485722b848
--- /dev/null
+++ b/whelktool/scripts/dataimports/shb/ConvertSHB.java
@@ -0,0 +1,33 @@
+import java.io.*;
+import java.util.*;
+import static java.lang.System.out;
+
+import se.kb.libris.util.marc.MarcRecord;
+import se.kb.libris.util.marc.io.Iso2709MarcRecordReader;
+
+import static whelk.util.Jackson.mapper;
+import whelk.converter.MarcJSONConverter;
+import whelk.converter.marc.MarcFrameConverter;
+import whelk.converter.marc.MarcFrameCli;
+
+// Example usage:
+// $ java -cp importers/build/libs/xlimporter.jar -Dxl.secret.properties=DEV2-secret.properties librisxl-tools/scripts/ConvertSHB.java ~/Downloads/shb_alfanum_marc > /var/tmp/shb_alnum.jsonld.lines
+public class ConvertSHB {
+  public static void main(String[] args) throws Exception {
+    var converter = new MarcFrameConverter();
+    new MarcFrameCli().addSystemComponents(converter);
+
+    var reader = new Iso2709MarcRecordReader(MarcJSONConverter.getNormalizedInputStreamFromFile(new File(args[0])));
+    MarcRecord record = null;
+    int i = 0;
+    while (true) {
+      record = reader.readRecord();
+      if (record == null)
+        break;
+
+      String id = "dataset/shb/" + (++i);
+      var result = (Map) converter.convert(MarcJSONConverter.toJSONMap(record), id);
+      out.println(mapper.writeValueAsString(result));
+    }
+  }
+}
diff --git a/whelktool/scripts/dataimports/shb/Makefile b/whelktool/scripts/dataimports/shb/Makefile
new file mode 100644
index 0000000000..e476868233
--- /dev/null
+++ b/whelktool/scripts/dataimports/shb/Makefile
@@ -0,0 +1,19 @@
+ROOT=../../../../
+STORE=$(HOME)/Downloads
+
+build/shb_alnum-reshaped.jsonld.lines: reshape_shb.py cache/shb_alnum-raw.jsonld.lines
+	mkdir -p build
+	python3 $^ cache/register_shb_240502-*.csv > $@
+
+cache/shb_alnum-raw.jsonld.lines: ConvertSHB.java $(STORE)/shb_alfanum_marc
+	mkdir -p cache
+	java -cp $(ROOT)importers/build/libs/xlimporter.jar -Dxl.secret.properties=$(ROOT)/DEV2-secret.properties $^ > $@
+
+build/shb-sample.trig: reshape_shb.py cache/shb_alnum-raw.jsonld.lines
+	mkdir -p build
+	python3 $^ cache/register_shb_240502-*.csv --sample-pretty-with $(ROOT)/../definitions/build/sys/context/kbv.jsonld | trld -ijsonld -o trig > $@
+
+# See <https://wiki.documentfoundation.org/ReleaseNotes/7.2#Document_Conversion>
+cache/register_shb_240502-*.csv: $(STORE)/register_shb_240502.xslx
+	mkdir -p cache
+	libreoffice --headless --convert-to csv:"Text - txt - csv (StarCalc)":"44,34,UTF8,1,,0,false,true,false,false,false,-1" $^ --outdir cache
diff --git a/whelktool/scripts/dataimports/shb/reshape_shb.py b/whelktool/scripts/dataimports/shb/reshape_shb.py
new file mode 100644
index 0000000000..ea01bf56d8
--- /dev/null
+++ b/whelktool/scripts/dataimports/shb/reshape_shb.py
@@ -0,0 +1,376 @@
+import csv
+import json
+import re
+import sys
+from pathlib import Path
+
+USE_ANNOT = True
+
+
+def convert(data, biblios: dict, subject_mappings) -> dict | None:
+    graph = data['@graph']
+    rec, thing, *rest = graph
+
+    if 'bibliography' in rec:
+        rec['bibliography'] = [{'@id': it['@id']} for it in rec['bibliography']]
+
+    del rec['marc:catalogingSource']  # "Annan verksamhet"
+    del rec['marc:typeOfControl']
+
+    if rest:
+        work = rest[0]
+        assert work['@id'].endswith('#work')
+        del work['@id']
+        del graph[2]
+    else:
+        work = {'@type': 'Work'}
+
+    thing['instanceOf'] = work
+
+    startyear: str | None = None
+    publ_year: str | None = None
+
+    if 'marc:primaryProvisionActivity' in thing and 'publication' in thing:
+        publ = thing['publication'][0]
+        startyear = publ.get('startYear')
+        publ_year = publ.get('endYear') or publ.get('year')
+        if publ_year == thing['marc:primaryProvisionActivity']['year']:
+            del thing['marc:primaryProvisionActivity']
+
+    partnum = thing.pop("part")[0]
+    iri: str | None = None
+    source: dict | None = None
+
+    if hosts := thing.pop('isPartOf'):
+        assert len(hosts) == 1
+        host = hosts[0]
+
+        if 'associatedMedia' in thing:
+            host['associatedMedia'] = [
+                {
+                    "@id": am['uri'][0].replace(
+                        'http://regina.kb.se/shb/', 'https://shb.kb.se/'
+                    )
+                }
+                for am in thing.pop('associatedMedia')
+            ]
+
+        if 'publication' in thing:
+            host['hasTemporalCoverage'] = thing.pop('publication')[0]
+            host['hasTemporalCoverage']['@type'] = 'TemporalCoverage'
+
+        hostrecs = host.pop('describedBy')
+        assert len(hostrecs) == 1
+
+        if ctrlnr := hostrecs[0].get('controlNumber'):
+            iri = f"http://libris.kb.se/resource/bib/{ctrlnr}"
+            host['@id'] = iri
+            if iri in biblios:
+                existing_host_desc = biblios[iri]
+                has_biblio_repr = json.dumps(existing_host_desc, sort_keys=True)
+                new_biblio_repr = json.dumps(host, sort_keys=True)
+                if has_biblio_repr != new_biblio_repr:
+                    bigslice = int(len(has_biblio_repr) * 0.8)
+                    assert (
+                        has_biblio_repr[:bigslice] == new_biblio_repr[:bigslice]
+                    ), f"Mismatch:\n{has_biblio_repr}\n{new_biblio_repr}"
+            else:
+                biblios[iri] = host
+
+            # Alternative to bibliography annotation:
+            # thing['cataloguedIn'] = {'@id': iri, "@annotation": {"part": thing.pop("part")}} # in the work...
+            for biblioref in rec['bibliography']:
+                bibliography_shb = 'https://libris.kb.se/library/SHB'
+                if biblioref['@id'] == bibliography_shb:
+                    source_id = f"{rec['@id']}#{partnum}"
+                    source = {
+                        "@id": source_id,
+                        "@type": "Source",
+                        "isPartOf": {'@id': iri},
+                        "item": partnum,
+                    }
+                    # biblios[source_id] = source
+                    rec['describes'] = source
+                    annot: dict = {"source": {"@id": source_id}}
+                    if USE_ANNOT:
+                        biblioref['@annotation'] = annot
+                    else:
+                        annot["_object"] = {'@id': bibliography_shb}
+                        rec.setdefault('_statementBy', {})['bibliography'] = annot
+                    break
+
+    if 'hasNote' in thing:
+        assert len(thing['hasNote']) == 1
+        note = thing.pop('hasNote')[0]['label']
+
+        note = note.replace('—', '-').replace(' ', ' ').replace('  ', ' ')
+
+        if note == 'TABORT':
+            return None
+
+        PARTOFMARK = ' - I:'
+        if PARTOFMARK in note:
+            note, partofnote = note.split(PARTOFMARK, 1)
+        else:
+            partofnote = None
+
+        name, title, subtitle, rest = parse_note(note)
+
+        thing['hasTitle'] = {"@type": "Title", "mainTitle": title}
+        if subtitle:
+            thing['hasTitle']['subtitle'] = subtitle
+
+        thing['responsibilityStatement'] = name
+        if rest:
+            thing['hasNote'] = [{"@type": "Note", "label": rest}]
+
+        if partofnote:
+            pname, ptitle, subtitle, prest = parse_note(partofnote, dotnote=False)
+            # assert not subtitle
+            issue = {
+                "@type": "Issue",
+                "label": ptitle,
+            }
+            if pname:
+                issue["responsibilityStatement"] = pname
+            if prest:
+                if USE_ANNOT:
+                    issue["@annotation"] = {"comment": prest}
+                else:
+                    raise NotImplementedError  # TODO
+
+            thing['partOf'] = issue
+
+        # TODO: remove 'ComponentPart', only re-add for obviously paginated,
+        # partOf:s and/or "newspaper reference"?
+
+    if partnum and publ_year:
+        years_key = f'{startyear}-{publ_year}' if startyear else publ_year
+        if rownummap := subject_mappings.get(years_key):
+            if subjectrefs := rownummap.get(partnum):  # TODO: opt + 'a' ...
+                work_subjects = work.setdefault('subject', [])
+                work_subjects += [{'@id': s} for s in subjectrefs]
+                if USE_ANNOT and iri:
+                    for s in work_subjects:
+                        s['@annotation'] = {"source": source}
+            # else:
+            #    print(f"{partnum} not in {list(rownummap)} for {years_key}", file=sys.stderr)
+
+    return {'@id': graph[0]['@id'], '@graph': data}
+
+
+def parse_note(note, dotnote=True):
+    """
+    >>> parse_note("Surname, G.-N., Anything")
+    ('Surname, G.-N.', 'Anything', None, None)
+
+    >>> parse_note("Anything. Surname, G.-N., Stuff.")
+    (None, 'Anything', None, 'Surname, G.-N., Stuff.')
+
+    >>> parse_note("Schuck, A., H. Schücks enka & Co. AB 150 år. [Stockholm.] Sthlm 1947, 28 s.")
+    ('Schuck, A.', 'H. Schücks enka & Co. AB 150 år', None, '[Stockholm.] Sthlm 1947, 28 s.')
+
+    >>> parse_note("Meyerson, Å., Ett besök vid Stora Kopparberget och Sala gruva år 1662. (BBV 23 (1938), s. 325-343.)")
+    ('Meyerson, Å.', 'Ett besök vid Stora Kopparberget och Sala gruva år 1662', None, 'BBV 23 (1938), s. 325-343.')
+
+    >>> parse_note('Davidsson, Åke, "En hoop Discantzböcker i godt förhwar...". Nyköping, 1976, s. 48-62')
+    ('Davidsson, Åke', '"En hoop Discantzböcker i godt förhwar..."', None, 'Nyköping, 1976, s. 48-62')
+
+    >>> parse_note('Davidsson, Åke, "En hoop Discantzböcker i godt förhwar..." : någotom Strängnäsgymnasiets musiksamling under 1600-talet. - I: Frånbiskop Rogge till Roggebiblioteket. Nyköping, 1976, s. 48-62')
+    ('Davidsson, Åke', '"En hoop Discantzböcker i godt förhwar..."', None, 'någotom Strängnäsgymnasiets musiksamling under 1600-talet. - I: Frånbiskop Rogge till Roggebiblioteket. Nyköping, 1976, s. 48-62')
+
+    >>> parse_note('The Swedish pioneer, ISSN0039-7326, 27, 1976:3, s. 215-221')
+    (None, 'The Swedish pioneer, ISSN0039-7326, 27, 1976:3', None, 's. 215-221')
+    """
+    ()
+    personnameparts = []
+    initial = -1
+    commaseparated = note.split(',')
+    name = commaseparated.pop(0).strip()
+    rest = ''
+
+    if ' ' in name.strip():
+        name = None
+        rest = note
+
+    elif commaseparated:
+        first = commaseparated[0]
+        if looks_like_initial(first.strip()) or re.match(r'^(\w|\s)+$', first):
+            name += ', ' + commaseparated.pop(0).strip()
+        else:
+            commaseparated.insert(0, name)
+            name = None
+
+        rest = ','.join(commaseparated).strip()
+
+    subtitle = ""
+
+    if ' : ' in rest:
+        title, rest = rest.split(' : ', 1)
+        if ' : ' in rest:
+            subtitle, rest = rest.split(' : ', 1)
+    elif rest.endswith(')'):
+        opencount = 0
+        for i, c in enumerate(rest[::-1]):
+            if c == ')':
+                opencount += 1
+            if c == '(':
+                opencount -= 1
+                if opencount == 0:
+                    break
+        title, rest = rest[: -i - 1], rest[-i:-1]
+        title = title.strip().removesuffix('.')
+    else:
+        x = int(rest.count('. ') * 0.3) or 1
+        title, *rest = rest.rsplit('. ', x) if dotnote and '. ' in rest else (rest, "")
+        rest = '. '.join(rest)
+        restispages = all(s.strip().isdigit() for s in rest.split('-'))
+        if len(title.strip()) == 1:
+            moretitle, rest = rest.split('.', 1) if '.' in rest else (rest, "")
+            title += '.' + moretitle
+        elif restispages:
+            title, prerest = title.rsplit('. ', 1) if '. ' in title else (title, "")
+            if not prerest:
+                title, posttitle = title.rsplit(' ', 1)
+                prerest = posttitle.strip()
+                title = title.strip().removesuffix(',')
+
+            rest = prerest + '. ' + rest
+
+    return name, title.strip(), subtitle.strip() or None, rest.strip() or None
+
+
+def looks_like_initial(s: str) -> bool:
+    """
+    >>> looks_like_initial("")
+    False
+    >>> looks_like_initial("A")
+    False
+    >>> looks_like_initial("A.")
+    True
+    >>> looks_like_initial("Ab.")
+    False
+    >>> looks_like_initial("A.-B.")
+    True
+    """
+    if not s:
+        return False
+    return s[0].isupper() and s.endswith('.') and (len(s) == 2 or s[-2].isupper())
+
+
+def pretty_print_sample(lines, biblios, subject_mappings, context_file):
+    with open(context_file) as f:
+        ctx = json.load(f)
+
+    results = []
+
+    x = 0
+    for i, l in enumerate(lines):
+        x += 1
+        if x > 19:
+            if i % 10_000 != 0:
+                continue
+            else:
+                x = 0
+        if data := convert(json.loads(l), biblios, subject_mappings):
+            results.append(data)
+
+    results = list(biblios.values()) + results
+
+    print(
+        json.dumps({'@context': ctx['@context'], '@graph': results}, ensure_ascii=False)
+    )
+
+
+def make_subject_mappings(subject_mapping_sheets: list[str]) -> dict:
+    subject_mappings: dict = {}
+
+    for sheet_file in subject_mapping_sheets:
+        _load_subject_mappings(subject_mappings, Path(sheet_file))
+
+    return subject_mappings
+
+
+def _load_subject_mappings(subject_mappings: dict, sheet_file: Path) -> None:
+    years_key = sheet_file.with_suffix('').name.split('-', 1)[-1]
+    assert years_key not in subject_mappings
+    rownummap = subject_mappings[years_key] = {}
+
+    with sheet_file.open() as f:
+        for i, row in enumerate(csv.reader(f)):
+            subjects: list[str] = []
+
+            startnum: str | None = None
+            endnum: str | None = None
+
+            for x in row[::-1]:
+                if not subjects and not x:
+                    continue
+
+                if x.startswith('https://id.kb.se/term/'):
+                    subjects.append(x)
+                elif endnum is None:
+                    endnum = x
+                elif startnum is None:
+                    startnum = x
+                else:
+                    break
+
+            subjects.reverse()
+
+            if subjects:
+                if not startnum:
+                    print(
+                        "Missing startnum in",
+                        sheet_file,
+                        "row",
+                        i,
+                        row,
+                        file=sys.stderr,
+                    )
+                    continue
+
+                assert startnum
+
+                startnum = startnum.strip()
+                if startnum.endswith('a'):
+                    startnum = startnum[:-1]
+
+                if endnum:
+                    endnum = endnum.strip()
+                    if endnum.endswith('a'):
+                        endnum = endnum[:-1]
+
+                    for n in range(int(startnum), int(endnum) + 1):
+                        rownummap[f"{n}"] = subjects
+                else:
+                    rownummap[f"{startnum}+"] = subjects
+
+
+if __name__ == '__main__':
+    import argparse
+
+    argp = argparse.ArgumentParser()
+    argp.add_argument('-t', '--test', action='store_true', default=False)
+    argp.add_argument('--sample-pretty-with')
+    argp.add_argument('infile')
+    argp.add_argument('subject_mapping_sheets', nargs='*')
+    args = argp.parse_args()
+
+    if args.test:
+        import doctest
+
+        doctest.testmod()
+        sys.exit(0)
+
+    subject_mappings = make_subject_mappings(args.subject_mapping_sheets)
+
+    biblios: dict = {}
+
+    with open(args.infile) as f:
+        if args.sample_pretty_with:
+            pretty_print_sample(f, biblios, subject_mappings, args.sample_pretty_with)
+            sys.exit()
+        for l in f:
+            if data := convert(json.loads(l), biblios, subject_mappings):
+                print(json.dumps(data, ensure_ascii=False))