Add doc for APPROX_COUNT_DISTINCT aggregate function

functions-and-operators/aggregate-group-by-functions.md

@@ -62,6 +62,32 @@
 
 Except for the `GROUP_CONCAT()` and `APPROX_PERCENTILE()` functions, all the preceding functions can serve as [Window functions](/functions-and-operators/window-functions.md).
 
++ `APPROX_COUNT_DISTINCT(expr, [expr...])`
+
+    The usage of this function is almost same with `COUNT(DISTINCT)` but returns approximate result. It uses `BJKST` algorithm and consumes less memory during simultaneous computation of cardinality for a large number of data sets whose cardinality has power law distribution. Moreover, it's very accurate for data sets with small cardinality and very efficient on CPU.
+
+    The following example shows how to use this function:
+
+    ```sql
+    DROP TABLE IF EXISTS t;
+    CREATE TABLE t(a INT, b INT, c INT);
+    INSERT INTO t VALUES(1, 1, 1), (2, 1, 1), (2, 2, 1), (3, 1, 1), (5, 1, 2), (5, 1, 2), (6, 1, 2), (7, 1, 2);
+    ```
+
+    ```sql
+    SELECT APPROX_COUNT_DISTINCT(a, b) FROM t GROUP BY c;
+    ```
+
+    ```sql
+    +-----------------------------+
+    | approx_count_distinct(a, b) |
+    +-----------------------------+
+    |                           3 |
+    |                           4 |
+    +-----------------------------+
+    2 rows in set (0.00 sec)
+    ```
+
 ## GROUP BY modifiers
 
 Starting from v7.4.0, the `GROUP BY` clause of TiDB supports the `WITH ROLLUP` modifier. For more information, see [GROUP BY modifiers](/functions-and-operators/group-by-modifier.md).