(doris) 20/23: [test](bloom_filter_func) add case about algorithm

[panxiaolei]

This is an automated email from the ASF dual-hosted git repository.

panxiaolei pushed a commit to branch refactor_rf
in repository https://gitbox.apache.org/repos/asf/doris.git

commit 4b6720a403335854262f5ee6ac606f599b8cc066
Author: Hu Shenggang <hushengg...@selectdb.com>
AuthorDate: Fri Mar 7 11:54:30 2025 +0800

    [test](bloom_filter_func) add case about algorithm
---
 be/test/exprs/bloom_filter_func_test.cpp | 58 ++++++++++++++++++++++++++++++++
 1 file changed, 58 insertions(+)

diff --git a/be/test/exprs/bloom_filter_func_test.cpp 
b/be/test/exprs/bloom_filter_func_test.cpp
index 5562132cc58..a99452ecda6 100644
--- a/be/test/exprs/bloom_filter_func_test.cpp
+++ b/be/test/exprs/bloom_filter_func_test.cpp
@@ -34,6 +34,7 @@
 #include "runtime/define_primitive_type.h"
 #include "runtime/primitive_type.h"
 #include "testutil/column_helper.h"
+#include "util/url_coding.h"
 #include "vec/columns/column_decimal.h"
 
 namespace doris {
@@ -314,6 +315,63 @@ TEST_F(BloomFilterFuncTest, Merge) {
     ASSERT_FALSE(st);
 }
 
+/// The purpose of this case is to detect changes after modifying the Bloom 
filter algorithm to prevent compatibility issues.
+TEST_F(BloomFilterFuncTest, HashAlgorithm) {
+    std::string BloomFilterBinary =
+            
"AAAAQAAAAIAACAAAAABAACAAAAAAQAAAAQAAAACAAAAAACAEAACAAgAACAQEAAAgAASAAAABAgAAEAAIDAAAAA"
+            
"CQAQAAYgAAAAIBAiBAgAABAgBABAAICAAABCGAAIABBAAAAAAAIABAAAAACAAAAAAAABAAQAAAAAAAIBAAAAAA"
+            
"AQAiABEAAQBAIgAAgBAQEEAAACACAQAABEgAAggAAQAAAUAQAAEQECCAAABAAIgHAAAACAEAAgAJQABAIAEAAA"
+            
"gAAEAAAAAAEAAAAAAQAAABAAAQAAAAAEAAAAAEAACAEAAAAAUAAAAAIBAgCAAAQAAIAAAACBAIABAAAAAABg";
+    BloomFilterFunc<PrimitiveType::TYPE_INT> bloom_filter_func(false);
+    const size_t runtime_length = 1024;
+    RuntimeFilterParams params {1,
+                                RuntimeFilterType::BLOOM_FILTER,
+                                PrimitiveType::TYPE_INT,
+                                false,
+                                0,
+                                0,
+                                0,
+                                256,
+                                0,
+                                0,
+                                false,
+                                false};
+    bloom_filter_func.init_params(&params);
+
+    ASSERT_TRUE(bloom_filter_func.init_with_fixed_length(runtime_length));
+
+    auto column = 
vectorized::ColumnHelper::create_column<vectorized::DataTypeInt32>(
+            {1, 3, 5, 7, 9, 12, 14, 16, 2001, 2002, 2003, 4096, 4097, 4098, 
4099, 4100});
+
+    bloom_filter_func.insert_fixed_len(column, 0);
+
+    char* data = nullptr;
+    int size;
+    bloom_filter_func.get_data(&data, &size);
+
+    std::string encode_string;
+    base64_encode(std::string(data, size), &encode_string);
+    ASSERT_EQ(strlen(BloomFilterBinary.c_str()), 
strlen(encode_string.c_str()));
+    ASSERT_EQ(memcmp(BloomFilterBinary.data(), encode_string.data(),
+                     strlen(BloomFilterBinary.c_str())),
+              0);
+
+    params.enable_fixed_len_to_uint32_v2 = true;
+
+    BloomFilterFunc<PrimitiveType::TYPE_INT> bloom_filter_func2(false);
+    bloom_filter_func2.init_params(&params);
+    ASSERT_TRUE(bloom_filter_func2.init_with_fixed_length(runtime_length));
+
+    bloom_filter_func2.insert_fixed_len(column, 0);
+    bloom_filter_func.get_data(&data, &size);
+    base64_encode(std::string(data, size), &encode_string);
+
+    ASSERT_EQ(strlen(BloomFilterBinary.c_str()), 
strlen(encode_string.c_str()));
+    ASSERT_EQ(memcmp(BloomFilterBinary.data(), encode_string.data(),
+                     strlen(BloomFilterBinary.c_str())),
+              0);
+}
+
 TEST_F(BloomFilterFuncTest, MergeLargeData) {
     BloomFilterFunc<PrimitiveType::TYPE_INT> bloom_filter_func(false);
     const size_t runtime_length = 1024;


---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscr...@doris.apache.org
For additional commands, e-mail: commits-h...@doris.apache.org