Added implementation of nunique function

docs/user-guide/advanced/Pandas_API.ipynb

@@ -2507,6 +2507,46 @@
     "tab.prod(numeric_only=True)"
    ]
   },
+  {
+   "cell_type": "markdown",
+   "source": [
+    "### Table.nunique()\n",
+    "```\n",
+    "Table.nunique(axis=0, skipna=True, numeric_only=False, min_count=0)\n",
+    "```\n",
+    "\n",
+    "Returns the number of unique elements across the given axis.\n",
+    "\n",
+    "**Parameters:**\n",
+    "\n",
+    "| Name         | Type | Description                                                                      | Default |\n",
+    "| :----------: | :--: | :------------------------------------------------------------------------------- | :-----: |\n",
+    "| axis         | int  | The axis to calculate the sum across 0 is columns, 1 is rows.                    | 0       |\n",
+    "| dropna       | bool | Don’t include NaN in the counts.                                                 | True    |\n",
+    "\n",
+    "**Returns:**\n",
+    "\n",
+    " | Type               | Description                                                          |\n",
+    " | :----------------: | :------------------------------------------------------------------- |\n",
+    " | Dictionary         | A dictionary where the key represent the column name / row number and the values are the result of calling `nunique` on that column / row. |"
+   ],
+   "metadata": {
+    "collapsed": false
+   },
+   "id": "7b39a07bd7cd0af7"
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "outputs": [],
+   "source": [
+    "tab.nunique()"
+   ],
+   "metadata": {
+    "collapsed": false
+   },
+   "id": "f5592b19b69ad46d"
+  },
   {
    "cell_type": "markdown",
    "id": "655c3ad2",

src/pykx/pandas_api/pandas_meta.py

@@ -255,3 +255,12 @@ def sum(self, axis=0, skipna=True, numeric_only=False, min_count=0):
             res,
             min_count
         ), cols)
+
+    @convert_result
+    def nunique(self, axis=0, dropna=True):
+        res, cols = preparse_computations(self, axis, skipna=False)
+        filterNan = q('{$[11h = type x;x;'
+                 '0h = type x;(x where not null x except w),(w:x where 10h=type each x);'
+                 'x where not null x]}each')
+        res = filterNan(res) if dropna else res
+        return q('(\'[count;distinct]\')', res), cols

tests/test_pandas_api.py

@@ -2029,3 +2029,40 @@ def test_keyed_loc_fixes(q):
         mkt[['k1', 'y']]
     with pytest.raises(KeyError):
         mkt['k1']
+
+
+def test_nunique(kx, q):
+    df = pd.DataFrame(
+        {
+            'a': [1, 2, 2, 4],
+            'b': [1, 2, 6, 7],
+            'c': [7, 8, 9, 10],
+            'd': ['foo', 'baz', 'baz', 'qux']
+        }
+    )
+    tab = kx.toq(df)
+    p_m = df.nunique()
+    q_m = tab.nunique()
+    for c in q.key(q_m).py():
+        assert p_m[c] == q_m[c].py()
+    p_m = df.nunique(axis=1)
+    q_m = tab.nunique(axis=1)
+    for c in range(len(tab)):
+        assert p_m[c] == q_m[c].py()
+
+    tab = kx.q('([]A:4 0n 7 6;B:4 0n 0n 7;C:``foo`foo`)')
+    df = tab.pd()
+    p_m = df.nunique()
+    q_m = tab.nunique()
+    for c in q.key(q_m).py():
+        assert p_m[c] == q_m[c].py()
+    p_m = df.nunique(axis=1, dropna=False)
+    q_m = tab.nunique(axis=1, dropna=False)
+    for c in range(len(tab)):
+        assert p_m[c] == q_m[c].py()
+    p_m = df.nunique(dropna=False)
+    q_m = tab.nunique(dropna=False)
+    for c in q.key(q_m).py():
+        assert p_m[c] == q_m[c].py()
+
+