bpo-39652: Truncate the column name after '[' only if PARSE_COLNAMES is set. (GH-18942)

serhiy-storchaka · web-flow · commit b146568dfcbc · 2020-03-21T15:53:28.000+02:00
diff --git a/Doc/library/sqlite3.rst b/Doc/library/sqlite3.rst
@@ -165,9 +165,10 @@ Module functions and constants
    that 'mytype' is the type of the column. It will try to find an entry of
    'mytype' in the converters dictionary and then use the converter function found
    there to return the value. The column name found in :attr:`Cursor.description`
-   is only the first word of the column name, i.  e. if you use something like
-   ``'as "x [datetime]"'`` in your SQL, then we will parse out everything until the
-   first blank for the column name: the column name would simply be "x".
+   does not include the type, i. e. if you use something like
+   ``'as "Expiration date [datetime]"'`` in your SQL, then we will parse out
+   everything until the first ``'['`` for the column name and strip
+   the preceeding space: the column name would simply be "Expiration date".
 
 
 .. function:: connect(database[, timeout, detect_types, isolation_level, check_same_thread, factory, cached_statements, uri])
diff --git a/Lib/sqlite3/test/regression.py b/Lib/sqlite3/test/regression.py
@@ -68,7 +68,7 @@ def CheckStatementReset(self):
     def CheckColumnNameWithSpaces(self):
         cur = self.con.cursor()
         cur.execute('select 1 as "foo bar [datetime]"')
-        self.assertEqual(cur.description[0][0], "foo bar")
+        self.assertEqual(cur.description[0][0], "foo bar [datetime]")
 
         cur.execute('select 1 as "foo baz"')
         self.assertEqual(cur.description[0][0], "foo baz")
diff --git a/Lib/sqlite3/test/types.py b/Lib/sqlite3/test/types.py
@@ -275,13 +275,13 @@ def CheckNone(self):
 
     def CheckColName(self):
         self.cur.execute("insert into test(x) values (?)", ("xxx",))
-        self.cur.execute('select x as "x [bar]" from test')
+        self.cur.execute('select x as "x y [bar]" from test')
         val = self.cur.fetchone()[0]
         self.assertEqual(val, "<xxx>")
 
         # Check if the stripping of colnames works. Everything after the first
-        # whitespace should be stripped.
-        self.assertEqual(self.cur.description[0][0], "x")
+        # '[' (and the preceeding space) should be stripped.
+        self.assertEqual(self.cur.description[0][0], "x y")
 
     def CheckCaseInConverterName(self):
         self.cur.execute("select 'other' as \"x [b1b1]\"")
diff --git a/Misc/NEWS.d/next/Library/2020-03-11-23-08-25.bpo-39652.gbasrk.rst b/Misc/NEWS.d/next/Library/2020-03-11-23-08-25.bpo-39652.gbasrk.rst
@@ -0,0 +1,2 @@
+The column name found in ``sqlite3.Cursor.description`` is now truncated on
+the first '[' only if the PARSE_COLNAMES option is set.
diff --git a/Modules/_sqlite/cursor.c b/Modules/_sqlite/cursor.c
@@ -193,22 +193,30 @@ pysqlite_build_row_cast_map(pysqlite_Cursor* self)
 }
 
 static PyObject *
-_pysqlite_build_column_name(const char* colname)
+_pysqlite_build_column_name(pysqlite_Cursor *self, const char *colname)
 {
     const char* pos;
+    Py_ssize_t len;
 
     if (!colname) {
         Py_RETURN_NONE;
     }
 
-    for (pos = colname;; pos++) {
-        if (*pos == 0 || *pos == '[') {
-            if ((*pos == '[') && (pos > colname) && (*(pos-1) == ' ')) {
-                pos--;
+    if (self->connection->detect_types & PARSE_COLNAMES) {
+        for (pos = colname; *pos; pos++) {
+            if (*pos == '[') {
+                if ((pos != colname) && (*(pos-1) == ' ')) {
+                    pos--;
+                }
+                break;
             }
-            return PyUnicode_FromStringAndSize(colname, pos - colname);
         }
+        len = pos - colname;
+    }
+    else {
+        len = strlen(colname);
     }
+    return PyUnicode_FromStringAndSize(colname, len);
 }
 
 /*
@@ -370,6 +378,7 @@ _pysqlite_query_execute(pysqlite_Cursor* self, int multiple, PyObject* args)
     PyObject* result;
     int numcols;
     PyObject* descriptor;
+    PyObject* column_name;
     PyObject* second_argument = NULL;
     sqlite_int64 lastrowid;
 
@@ -536,7 +545,13 @@ _pysqlite_query_execute(pysqlite_Cursor* self, int multiple, PyObject* args)
                 if (!descriptor) {
                     goto error;
                 }
-                PyTuple_SetItem(descriptor, 0, _pysqlite_build_column_name(sqlite3_column_name(self->statement->st, i)));
+                column_name = _pysqlite_build_column_name(self,
+                                sqlite3_column_name(self->statement->st, i));
+                if (!column_name) {
+                    Py_DECREF(descriptor);
+                    goto error;
+                }
+                PyTuple_SetItem(descriptor, 0, column_name);
                 Py_INCREF(Py_None); PyTuple_SetItem(descriptor, 1, Py_None);
                 Py_INCREF(Py_None); PyTuple_SetItem(descriptor, 2, Py_None);
                 Py_INCREF(Py_None); PyTuple_SetItem(descriptor, 3, Py_None);

Original file line number	Diff line number	Diff line change
`@@ -0,0 +1,2 @@`
	`1`	+The column name found in ``sqlite3.Cursor.description`` is now truncated on
	`2`	`+the first '[' only if the PARSE_COLNAMES option is set.`