#26545 Fix: same .tsv file, get different data-frame structure using engine 'python' and 'c'

doc/source/whatsnew/v0.25.0.rst

@@ -667,6 +667,7 @@ I/O
 - Bug in :func:`read_json` where date strings with ``Z`` were not converted to a UTC timezone (:issue:`26168`)
 - Added ``cache_dates=True`` parameter to :meth:`read_csv`, which allows to cache unique dates when they are parsed (:issue:`25990`)
 - :meth:`DataFrame.to_excel` now raises a ``ValueError`` when the caller's dimensions exceed the limitations of Excel (:issue:`26051`)
+- Fixed bug in :func:`pandas.read_csv` where a BOM would result in incorrect parsing using engine='python' (:issue:`26545`)
 - :func:`read_excel` now raises a ``ValueError`` when input is of type :class:`pandas.io.excel.ExcelFile` and ``engine`` param is passed since :class:`pandas.io.excel.ExcelFile` has an engine defined (:issue:`26566`)
 - Bug while selecting from :class:`HDFStore` with ``where=''`` specified (:issue:`26610`).
 

pandas/io/parsers.py

@@ -2755,23 +2755,24 @@ def _check_for_bom(self, first_row):
         if first_elt != _BOM:
             return first_row
 
-        first_row = first_row[0]
+        first_row_bom = first_row[0]
 
-        if len(first_row) > 1 and first_row[1] == self.quotechar:
+        if len(first_row_bom) > 1 and first_row_bom[1] == self.quotechar:
             start = 2
-            quote = first_row[1]
-            end = first_row[2:].index(quote) + 2
+            quote = first_row_bom[1]
+            end = first_row_bom[2:].index(quote) + 2
 
             # Extract the data between the quotation marks
-            new_row = first_row[start:end]
+            new_row = first_row_bom[start:end]
 
             # Extract any remaining data after the second
             # quotation mark.
-            if len(first_row) > end + 1:
-                new_row += first_row[end + 1:]
-            return [new_row]
-        elif len(first_row) > 1:
-            return [first_row[1:]]
+            if len(first_row_bom) > end + 1:
+                new_row += first_row_bom[end + 1:]
+            return [new_row] + first_row[1:]
+
+        elif len(first_row_bom) > 1:
+            return [first_row_bom[1:]]
         else:
             # First row is just the BOM, so we
             # return an empty string.

pandas/tests/io/parser/test_common.py

@@ -1927,3 +1927,13 @@ def test_read_table_deprecated(all_parsers):
                                     check_stacklevel=False):
         result = parser.read_table(StringIO(data))
         tm.assert_frame_equal(result, expected)
+
+
+def test_first_row_bom(all_parsers):
+    # see gh-26545
+    parser = all_parsers
+    data = '''\ufeff"Head1"	"Head2"	"Head3"'''
+
+    result = parser.read_csv(StringIO(data), delimiter='\t')
+    expected = DataFrame(columns=["Head1", "Head2", "Head3"])
+    tm.assert_frame_equal(result, expected)