From 5f44ce8c40795324cc2b841af858319a742119de Mon Sep 17 00:00:00 2001
From: Seaven <seaven_7@qq.com>
Date: Fri, 24 Jan 2025 19:05:27 +0800
Subject: [PATCH] [Enhancement] Decode function add length  (#55315)

Signed-off-by: Seaven <seaven_7@qq.com>
---
 .../rule/tree/lowcardinality/DecodeCollector.java  |  2 +-
 .../starrocks/sql/plan/LowCardinalityTest2.java    |  9 +++++++++
 .../test_low_cardinality/R/test_low_cardinality2   | 14 +++++++++-----
 .../test_low_cardinality/T/test_low_cardinality2   |  1 +
 4 files changed, 20 insertions(+), 6 deletions(-)

diff --git a/fe/fe-core/src/main/java/com/starrocks/sql/optimizer/rule/tree/lowcardinality/DecodeCollector.java b/fe/fe-core/src/main/java/com/starrocks/sql/optimizer/rule/tree/lowcardinality/DecodeCollector.java
index 160a87704d9cd6..0f5964d1412fc8 100644
--- a/fe/fe-core/src/main/java/com/starrocks/sql/optimizer/rule/tree/lowcardinality/DecodeCollector.java
+++ b/fe/fe-core/src/main/java/com/starrocks/sql/optimizer/rule/tree/lowcardinality/DecodeCollector.java
@@ -91,7 +91,7 @@ public class DecodeCollector extends OptExpressionVisitor<DecodeInfo, DecodeInfo
                     FunctionSet.LTRIM, FunctionSet.REGEXP_EXTRACT, FunctionSet.REGEXP_REPLACE, FunctionSet.REPEAT,
                     FunctionSet.REPLACE, FunctionSet.REVERSE, FunctionSet.RIGHT, FunctionSet.RPAD, FunctionSet.RTRIM,
                     FunctionSet.SPLIT_PART, FunctionSet.SUBSTR, FunctionSet.SUBSTRING, FunctionSet.SUBSTRING_INDEX,
-                    FunctionSet.TRIM, FunctionSet.UPPER, FunctionSet.IF);
+                    FunctionSet.TRIM, FunctionSet.UPPER, FunctionSet.IF, FunctionSet.LENGTH, FunctionSet.CHAR_LENGTH);
 
     // array<string> support:
     //  array<string> -> array<string>: array function
diff --git a/fe/fe-core/src/test/java/com/starrocks/sql/plan/LowCardinalityTest2.java b/fe/fe-core/src/test/java/com/starrocks/sql/plan/LowCardinalityTest2.java
index e2d0685907c106..adf84075c26352 100644
--- a/fe/fe-core/src/test/java/com/starrocks/sql/plan/LowCardinalityTest2.java
+++ b/fe/fe-core/src/test/java/com/starrocks/sql/plan/LowCardinalityTest2.java
@@ -304,6 +304,15 @@ public void testDecodeNodeRewrite4() throws Exception {
                 "  |  <slot 4> : 4: dept_name"));
     }
 
+    @Test
+    public void testDecodeNodeRewriteLength() throws Exception {
+        String sql = "select length(dept_name), char_length(dept_name) from dept group by dept_name,state";
+        String plan = getFragmentPlan(sql);
+        Assert.assertTrue(plan, plan.contains("  2:Project\n" +
+                "  |  <slot 4> : DictDecode(6: dept_name, [length(<place-holder>)])\n" +
+                "  |  <slot 5> : DictDecode(6: dept_name, [char_length(<place-holder>)])"));
+    }
+
     @Test
     public void testDecodeNodeRewrite5() throws Exception {
         String sql = "select S_ADDRESS from supplier where S_ADDRESS " +
diff --git a/test/sql/test_low_cardinality/R/test_low_cardinality2 b/test/sql/test_low_cardinality/R/test_low_cardinality2
index c75b6b5ac7f536..2a33193dd7f2cd 100644
--- a/test/sql/test_low_cardinality/R/test_low_cardinality2
+++ b/test/sql/test_low_cardinality/R/test_low_cardinality2
@@ -307,11 +307,11 @@ insert into s2 select * from s2;
 -- !result
 [UC] analyze full table s1;
 -- result:
-test_db_597f05742dee11efbc9b7b0d31512d80.s1	analyze	status	OK
+test_db_3569e679a63f46fcb95f5e935f2b4716.s1	analyze	status	OK
 -- !result
 [UC] analyze full table s2;
 -- result:
-test_db_597f05742dee11efbc9b7b0d31512d80.s2	analyze	status	OK
+test_db_3569e679a63f46fcb95f5e935f2b4716.s2	analyze	status	OK
 -- !result
 function: wait_global_dict_ready('v3', 's1')
 -- result:
@@ -338,6 +338,10 @@ select concat(upper(v4), "1"), max(v3) from s2 where v4 = "BJ" group by upper(v4
 -- result:
 BJ1	Jiangxi
 -- !result
+select length(v4), char_length(v4), max(v3) from s2 where v4 = "BJ" group by length(v4), char_length(v4) order by 1, 2;
+-- result:
+2	2	Jiangxi
+-- !result
 select count(distinct v3), max(v3), count(distinct v4) as a from s2;
 -- result:
 33	Zhejiang	34
@@ -484,7 +488,7 @@ FROM
 -- !result
 [UC] analyze full table supplier;
 -- result:
-test_db_597f05742dee11efbc9b7b0d31512d80.supplier	analyze	status	OK
+test_db_3569e679a63f46fcb95f5e935f2b4716.supplier	analyze	status	OK
 -- !result
 function: wait_global_dict_ready('s_region', 'supplier')
 -- result:
@@ -508,7 +512,7 @@ select l.S_NAME,l.mx_addr from agged_supplier_5 l right join [shuffle] supplier
 3	3
 4	4
 -- !result
- with a as ( select s_region, S_NAME from supplier where S_SUPPKEY < 10 limit 10 ), b as ( select max(upper(s_region)) as mx from a group by S_NAME ), c as ( select lower(mx) = lower('lw') from b )select *from c;
+with a as ( select s_region, S_NAME from supplier where S_SUPPKEY < 10 limit 10 ), b as ( select max(upper(s_region)) as mx from a group by S_NAME ), c as ( select lower(mx) = lower('lw') from b )select *from c;
 -- result:
 0
 0
@@ -519,4 +523,4 @@ select l.S_NAME,l.mx_addr from agged_supplier_5 l right join [shuffle] supplier
 0
 0
 0
--- !result
\ No newline at end of file
+-- !result
diff --git a/test/sql/test_low_cardinality/T/test_low_cardinality2 b/test/sql/test_low_cardinality/T/test_low_cardinality2
index f1ff12ec57d4ce..a75656da425262 100644
--- a/test/sql/test_low_cardinality/T/test_low_cardinality2
+++ b/test/sql/test_low_cardinality/T/test_low_cardinality2
@@ -304,6 +304,7 @@ function: wait_global_dict_ready('v4', 's2')
 
 select concat(upper(v2), "1"), max(v3) from s2 where v4 = "BJ" group by upper(v2) order by 1, 2;
 select concat(upper(v4), "1"), max(v3) from s2 where v4 = "BJ" group by upper(v4) order by 1, 2;
+select length(v4), char_length(v4), max(v3) from s2 where v4 = "BJ" group by length(v4), char_length(v4) order by 1, 2;
 
 select count(distinct v3), max(v3), count(distinct v4) as a from s2;