JSON Support for parsing NaN, Infinity and -Infinity (#30295)

Author: WillAyd

doc/source/whatsnew/v1.0.0.rst

@@ -216,6 +216,7 @@ Other enhancements
   (:meth:`~DataFrame.to_parquet` / :func:`read_parquet`) using the `'pyarrow'` engine
   now preserve those data types with pyarrow >= 1.0.0 (:issue:`20612`).
 - The ``partition_cols`` argument in :meth:`DataFrame.to_parquet` now accepts a string (:issue:`27117`)
+- :func:`pandas.read_json` now parses ``NaN``, ``Infinity`` and ``-Infinity`` (:issue:`12213`)
 - The ``pandas.np`` submodule is now deprecated. Import numpy directly instead (:issue:`30296`)
 - :func:`to_parquet` now appropriately handles the ``schema`` argument for user defined schemas in the pyarrow engine. (:issue: `30270`)
 - DataFrame constructor preserve `ExtensionArray` dtype with `ExtensionArray` (:issue:`11363`)

pandas/_libs/src/ujson/lib/ultrajson.h

@@ -154,6 +154,8 @@ enum JSTYPES {
   JT_ARRAY,    // Array structure
   JT_OBJECT,   // Key/Value structure
   JT_INVALID,  // Internal, do not return nor expect
+  JT_POS_INF,  // Positive infinity
+  JT_NEG_INF,  // Negative infinity
 };
 
 typedef void * JSOBJ;
@@ -290,6 +292,8 @@ typedef struct __JSONObjectDecoder {
   JSOBJ (*newTrue)(void *prv);
   JSOBJ (*newFalse)(void *prv);
   JSOBJ (*newNull)(void *prv);
+  JSOBJ (*newPosInf)(void *prv);
+  JSOBJ (*newNegInf)(void *prv);
   JSOBJ (*newObject)(void *prv, void *decoder);
   JSOBJ (*endObject)(void *prv, JSOBJ obj);
   JSOBJ (*newArray)(void *prv, void *decoder);

pandas/_libs/src/ujson/lib/ultrajsondec.c

@@ -127,9 +127,16 @@ FASTCALL_ATTR JSOBJ FASTCALL_MSVC decode_numeric(struct DecoderState *ds) {
 
     JSUINT64 overflowLimit = LLONG_MAX;
 
-    if (*(offset) == '-') {
+    if (*(offset) == 'I') {
+      goto DECODE_INF;
+    } else if (*(offset) == 'N') {
+      goto DECODE_NAN;
+    } else if (*(offset) == '-') {
         offset++;
         intNeg = -1;
+        if (*(offset) == 'I') {
+          goto DECODE_INF;
+        }
         overflowLimit = LLONG_MIN;
     }
 
@@ -281,6 +288,48 @@ FASTCALL_ATTR JSOBJ FASTCALL_MSVC decode_numeric(struct DecoderState *ds) {
         }
     }
 
+DECODE_NAN:
+    offset++;
+    if (*(offset++) != 'a') goto SET_NAN_ERROR;
+    if (*(offset++) != 'N') goto SET_NAN_ERROR;
+
+    ds->lastType = JT_NULL;
+    ds->start = offset;
+    return ds->dec->newNull(ds->prv);
+
+SET_NAN_ERROR:
+    return SetError(ds, -1, "Unexpected character found when decoding 'NaN'");
+
+DECODE_INF:
+    offset++;
+    if (*(offset++) != 'n') goto SET_INF_ERROR;
+    if (*(offset++) != 'f') goto SET_INF_ERROR;
+    if (*(offset++) != 'i') goto SET_INF_ERROR;
+    if (*(offset++) != 'n') goto SET_INF_ERROR;
+    if (*(offset++) != 'i') goto SET_INF_ERROR;
+    if (*(offset++) != 't') goto SET_INF_ERROR;
+    if (*(offset++) != 'y') goto SET_INF_ERROR;
+
+    ds->start = offset;
+
+    if (intNeg == 1) {
+      ds->lastType = JT_POS_INF;
+      return ds->dec->newPosInf(ds->prv);
+    } else {
+      ds->lastType = JT_NEG_INF;
+      return ds->dec->newNegInf(ds->prv);
+    }
+
+SET_INF_ERROR:
+    if (intNeg == 1) {
+      const char *msg = "Unexpected character found when decoding 'Infinity'";
+      return SetError(ds, -1, msg);
+    } else {
+      const char *msg = "Unexpected character found when decoding '-Infinity'";
+      return SetError(ds, -1, msg);
+    }
+
+
 BREAK_EXP_LOOP:
     // FIXME: Check for arithmetic overflow here
     ds->lastType = JT_DOUBLE;
@@ -1070,6 +1119,8 @@ FASTCALL_ATTR JSOBJ FASTCALL_MSVC decode_any(struct DecoderState *ds) {
             case '7':
             case '8':
             case '9':
+            case 'I':
+            case 'N':
             case '-':
                 return decode_numeric(ds);
 

pandas/_libs/src/ujson/python/JSONtoObj.c

@@ -459,6 +459,10 @@ JSOBJ Object_newFalse(void *prv) { Py_RETURN_FALSE; }
 
 JSOBJ Object_newNull(void *prv) { Py_RETURN_NONE; }
 
+JSOBJ Object_newPosInf(void *prv) { return PyFloat_FromDouble(Py_HUGE_VAL); }
+
+JSOBJ Object_newNegInf(void *prv) { return PyFloat_FromDouble(-Py_HUGE_VAL); }
+
 JSOBJ Object_newObject(void *prv, void *decoder) { return PyDict_New(); }
 
 JSOBJ Object_endObject(void *prv, JSOBJ obj) { return obj; }
@@ -502,10 +506,11 @@ PyObject *JSONToObj(PyObject *self, PyObject *args, PyObject *kwargs) {
     JSONObjectDecoder dec = {
         Object_newString, Object_objectAddKey,  Object_arrayAddItem,
         Object_newTrue,   Object_newFalse,      Object_newNull,
-        Object_newObject, Object_endObject,     Object_newArray,
-        Object_endArray,  Object_newInteger,    Object_newLong,
-        Object_newDouble, Object_releaseObject, PyObject_Malloc,
-        PyObject_Free,    PyObject_Realloc};
+        Object_newPosInf, Object_newNegInf,     Object_newObject,
+        Object_endObject,     Object_newArray,  Object_endArray,
+        Object_newInteger,    Object_newLong,   Object_newDouble,
+        Object_releaseObject, PyObject_Malloc, PyObject_Free,
+        PyObject_Realloc};
 
     dec.preciseFloat = 0;
     dec.prv = NULL;

pandas/tests/io/json/test_pandas.py

@@ -1597,3 +1597,12 @@ def test_json_indent_all_orients(self, orient, expected):
     def test_json_negative_indent_raises(self):
         with pytest.raises(ValueError, match="must be a nonnegative integer"):
             pd.DataFrame().to_json(indent=-1)
+
+    def test_emca_262_nan_inf_support(self):
+        # GH 12213
+        data = '["a", NaN, "NaN", Infinity, "Infinity", -Infinity, "-Infinity"]'
+        result = pd.read_json(data)
+        expected = pd.DataFrame(
+            ["a", np.nan, "NaN", np.inf, "Infinity", -np.inf, "-Infinity"]
+        )
+        tm.assert_frame_equal(result, expected)