fix(salary): never take a qualified salary header as the company column (#700)

* fix(salary): never take a qualified salary header as the company column

The detection loop kept the LAST header matching a company pattern, so
"Company Name | City | Company Count | Company Index" made the index
column the company column and every converted row carried
"company": "108.5"; IDA's "Firmanavn og by | Antal | Firmaets indeks"
(#615) did the same. #571's exact-match preference only helped when a
header was exactly "Company" or "Firma".

Detection is now three tiers, each taking the first match: an exact
pattern header, then a pattern header that is not a count/index column,
then any pattern header; the city column can never alias the company
column. Four new cases; the two qualified-header ones fail on master.

* changelog: place the #700 entry under [Unreleased]

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Mads Lorentzen <madslorentzen_17@hotmail.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

tools/convert_salary_excel.py

@@ -197,20 +197,28 @@ def parse_sheet(ws, sheet_label=None):
     for cell in ws[header_row]:
         headers.append(" ".join(str(cell.value).split()) if cell.value else "")
 
-    # Find company and city columns
-    company_col = None
-    city_col = None
-    for i, h in enumerate(headers):
-        if header_matches(h, COMPANY_PATTERNS):
-            company_col = i
-        elif header_matches(h, CITY_PATTERNS):
-            city_col = i
+    # Find company and city columns. Three tiers, each taking the FIRST match:
+    #   1. an exact pattern header ("Company", "Firma", "By");
+    #   2. a header that carries the pattern but is not a count/index column
+    #      ("Company Name", "Firmanavn og by");
+    #   3. any header carrying the pattern.
+    # The old loop kept the LAST pattern match, so "Company Name | ... |
+    # Company Index" made the salary index the company column: every row's
+    # company read "108.5". #571 added tier 1; tiers 2-3 cover a qualified
+    # name header that sits next to a qualified salary header. The city
+    # lookup skips the company column: a combined "Company and city" header
+    # carries both words, and must not shadow a real city column after it.
+    def metadata_column(patterns, skip=None):
+        usable = [(i, h) for i, h in enumerate(headers) if i != skip]
+        exact = next((i for i, h in usable if h.lower().strip() in patterns), None)
+        if exact is not None:
+            return exact
+        candidates = [i for i, h in usable if header_matches(h, patterns)]
+        plain = [i for i in candidates if detect_column_type(headers[i]) is None]
+        return (plain or candidates or [None])[0]
 
-    # A qualified salary header such as "Company Index" may also contain a
-    # metadata token. Prefer an unqualified metadata column when one exists,
-    # leaving the qualified column available as salary data.
-    company_col = next((i for i, h in enumerate(headers) if h.lower() in COMPANY_PATTERNS), company_col)
-    city_col = next((i for i, h in enumerate(headers) if h.lower() in CITY_PATTERNS), city_col)
+    company_col = metadata_column(COMPANY_PATTERNS)
+    city_col = metadata_column(CITY_PATTERNS, skip=company_col)
 
     if company_col is None:
         print(f"Warning: Could not find company column in sheet '{ws.title}'.", file=sys.stderr)