add deserialize test

mapleFU · Nov 24, 2023 · e5875e9 · e5875e9
1 parent f7535cc
commit e5875e9
Show file tree

Hide file tree

Showing 2 changed files with 25 additions and 19 deletions.
diff --git a/cpp/src/parquet/bloom_filter_reader.cc b/cpp/src/parquet/bloom_filter_reader.cc
@@ -64,10 +64,14 @@ std::unique_ptr<BloomFilter> RowGroupBloomFilterReaderImpl::GetColumnBloomFilter
     throw ParquetException("file size less or equal than bloom offset");
   }
   std::optional<int64_t> bloom_filter_length = col_chunk->bloom_filter_length();
-  if (bloom_filter_length.has_value() &&
-      *bloom_filter_length + *bloom_filter_offset > file_size) {
-    throw ParquetException(
-        "bloom filter length + bloom filter offset greater than file size");
+  if (bloom_filter_length.has_value()) {
+    if (*bloom_filter_length < 0) {
+      throw ParquetException("bloom_filter_length less than 0");
+    }
+    if (*bloom_filter_length + *bloom_filter_offset > file_size) {
+      throw ParquetException(
+          "bloom filter length + bloom filter offset greater than file size");
+    }
   }
   auto stream = ::arrow::io::RandomAccessFile::GetStream(
       input_, *bloom_filter_offset, file_size - *bloom_filter_offset);

diff --git a/cpp/src/parquet/bloom_filter_test.cc b/cpp/src/parquet/bloom_filter_test.cc
@@ -107,24 +107,26 @@ TEST(BasicTest, TestBloomFilter) {
 
     // Deserialize Bloom filter from memory
     ASSERT_OK_AND_ASSIGN(auto buffer, sink->Finish());
-    ::arrow::io::BufferReader source(buffer);
 
     ReaderProperties reader_properties;
-    BlockSplitBloomFilter de_bloom =
-        BlockSplitBloomFilter::Deserialize(reader_properties, &source, std::nullopt);
-
-    // Lookup previously inserted values
-    for (const auto v : kIntInserts) {
-      EXPECT_TRUE(de_bloom.FindHash(de_bloom.Hash(v)));
-    }
-    for (const auto v : kFloatInserts) {
-      EXPECT_TRUE(de_bloom.FindHash(de_bloom.Hash(v)));
+    for (std::optional<int64_t> bloom_filter_length :
+         std::vector<std::optional<int64_t>>{std::nullopt, buffer->size()}) {
+      ::arrow::io::BufferReader source(buffer);
+      BlockSplitBloomFilter de_bloom = BlockSplitBloomFilter::Deserialize(
+          reader_properties, &source, bloom_filter_length);
+      // Lookup previously inserted values
+      for (const auto v : kIntInserts) {
+        EXPECT_TRUE(de_bloom.FindHash(de_bloom.Hash(v)));
+      }
+      for (const auto v : kFloatInserts) {
+        EXPECT_TRUE(de_bloom.FindHash(de_bloom.Hash(v)));
+      }
+      false_positives = 0;
+      for (const auto v : kNegativeIntLookups) {
+        false_positives += de_bloom.FindHash(de_bloom.Hash(v));
+      }
+      EXPECT_LE(false_positives, 2);
     }
-    false_positives = 0;
-    for (const auto v : kNegativeIntLookups) {
-      false_positives += de_bloom.FindHash(de_bloom.Hash(v));
-    }
-    EXPECT_LE(false_positives, 2);
   }
 }