fix: pass dtypes to read_json with pyarrow engine

Author: will-larkin

doc/source/whatsnew/v3.0.0.rst

@@ -732,6 +732,7 @@ I/O
 - Bug in :meth:`read_stata` where the missing code for double was not recognised for format versions 105 and prior (:issue:`58149`)
 - Bug in :meth:`set_option` where setting the pandas option ``display.html.use_mathjax`` to ``False`` has no effect (:issue:`59884`)
 - Bug in :meth:`to_excel` where :class:`MultiIndex` columns would be merged to a single row when ``merge_cells=False`` is passed (:issue:`60274`)
+- Bug in :meth:`read_json` ignoring the given ``dtype`` when ``engine="pyarrow"`` (:issue:`59516`)
 
 Period
 ^^^^^^

pandas/io/json/_json.py

@@ -32,6 +32,7 @@
 from pandas.core.dtypes.common import (
     ensure_str,
     is_string_dtype,
+    pandas_dtype,
 )
 from pandas.core.dtypes.dtypes import PeriodDtype
 
@@ -43,6 +44,7 @@
     isna,
     notna,
     to_datetime,
+    ArrowDtype,
 )
 from pandas.core.reshape.concat import concat
 from pandas.core.shared_docs import _shared_docs
@@ -942,29 +944,56 @@ def read(self) -> DataFrame | Series:
         obj: DataFrame | Series
         with self:
             if self.engine == "pyarrow":
-                pyarrow_json = import_optional_dependency("pyarrow.json")
-                pa_table = pyarrow_json.read_json(self.data)
-                return arrow_table_to_pandas(pa_table, dtype_backend=self.dtype_backend)
+                obj = self._read_pyarrow()
             elif self.engine == "ujson":
-                if self.lines:
-                    if self.chunksize:
-                        obj = concat(self)
-                    elif self.nrows:
-                        lines = list(islice(self.data, self.nrows))
-                        lines_json = self._combine_lines(lines)
-                        obj = self._get_object_parser(lines_json)
-                    else:
-                        data = ensure_str(self.data)
-                        data_lines = data.split("\n")
-                        obj = self._get_object_parser(self._combine_lines(data_lines))
-                else:
-                    obj = self._get_object_parser(self.data)
-                if self.dtype_backend is not lib.no_default:
-                    return obj.convert_dtypes(
-                        infer_objects=False, dtype_backend=self.dtype_backend
-                    )
-                else:
-                    return obj
+                obj = self._read_ujson()
+        
+        return obj
+
+    def _read_pyarrow(self) -> DataFrame:
+        """
+        Read JSON using the pyarrow engine.
+        """
+        pyarrow_json = import_optional_dependency("pyarrow.json")
+        options = None
+
+        if isinstance(self.dtype, dict):
+            pa = import_optional_dependency("pyarrow")
+            fields = [
+                (field, pandas_dtype(dtype).pyarrow_dtype)
+                for field, dtype in self.dtype.items()
+                if isinstance(pandas_dtype(dtype), ArrowDtype)
+            ]
+
+            schema = pa.schema(fields)
+            options = pyarrow_json.ParseOptions(explicit_schema=schema)
+            
+        pa_table = pyarrow_json.read_json(self.data, parse_options=options)
+        return arrow_table_to_pandas(pa_table, dtype_backend=self.dtype_backend)
+    
+    def _read_ujson(self) -> DataFrame | Series:
+        """
+        Read JSON using the ujson engine.
+        """
+        if self.lines:
+            if self.chunksize:
+                obj = concat(self)
+            elif self.nrows:
+                lines = list(islice(self.data, self.nrows))
+                lines_json = self._combine_lines(lines)
+                obj = self._get_object_parser(lines_json)
+            else:
+                data = ensure_str(self.data)
+                data_lines = data.split("\n")
+                obj = self._get_object_parser(self._combine_lines(data_lines))
+        else:
+            obj = self._get_object_parser(self.data)
+        if self.dtype_backend is not lib.no_default:
+            return obj.convert_dtypes(
+                infer_objects=False, dtype_backend=self.dtype_backend
+            )
+        else:
+            return obj
 
     def _get_object_parser(self, json: str) -> DataFrame | Series:
         """

pandas/tests/io/json/test_pandas.py

@@ -2183,6 +2183,28 @@ def test_read_json_dtype_backend(
         # string_storage setting -> ignore that for checking the result
         tm.assert_frame_equal(result, expected, check_column_type=False)
 
+    @td.skip_if_no("pyarrow") # type: ignore
+    def test_read_json_pyarrow_with_dtype(self, datapath):
+        dtype = {"a": "int32[pyarrow]", "b": "int64[pyarrow]"}
+
+        df = read_json(
+            datapath("io", "json", "data", "line_delimited.json"),
+            dtype=dtype,
+            lines=True,
+            engine="pyarrow",
+            dtype_backend="pyarrow",
+        )
+
+        result = df.dtypes
+        expected = Series(
+            [
+                pd.ArrowDtype.construct_from_string("int32[pyarrow]"),
+                pd.ArrowDtype.construct_from_string("int64[pyarrow]"),
+            ],
+            index=["a", "b"],
+        )
+        tm.assert_series_equal(result, expected)
+
     @pytest.mark.parametrize("orient", ["split", "records", "index"])
     def test_read_json_nullable_series(self, string_storage, dtype_backend, orient):
         # GH#50750