Add list look up test (#3210)

* list look up test
kuzudb · Apr 7, 2024 · f391b28 · f391b28
1 parent c6897b4
commit f391b28
Show file tree

Hide file tree

Showing 3 changed files with 96 additions and 2 deletions.
diff --git a/src/storage/store/list_column.cpp b/src/storage/store/list_column.cpp
@@ -76,6 +76,7 @@ void ListColumn::scan(Transaction* transaction, node_group_idx_t nodeGroupIdx,
                                   resultVector->getValue<list_entry_t>(offsetInVector - 1).size;
     auto offsetToWriteListData = listOffsetInVector;
     auto numValues = endOffsetInGroup - startOffsetInGroup;
+    numValues = std::min(numValues, listOffsetInfoInStorage.numTotal);
     KU_ASSERT(numValues >= 0);
     for (auto i = 0u; i < numValues; i++) {
         list_size_t size = listOffsetInfoInStorage.getListSize(i);
@@ -197,6 +198,7 @@ void ListColumn::append(ColumnChunk* columnChunk, uint64_t nodeGroupIdx) {
 void ListColumn::scanUnfiltered(Transaction* transaction, node_group_idx_t nodeGroupIdx,
     ValueVector* resultVector, const ListOffsetSizeInfo& listOffsetInfoInStorage) {
     auto numValuesToScan = resultVector->state->selVector->selectedSize;
+    numValuesToScan = std::min(numValuesToScan, listOffsetInfoInStorage.numTotal);
     offset_t offsetInVector = 0;
     for (auto i = 0u; i < numValuesToScan; i++) {
         auto listLen = listOffsetInfoInStorage.getListSize(i);
@@ -303,7 +305,8 @@ ListOffsetSizeInfo ListColumn::getListOffsetSizeInfo(Transaction* transaction,
         endOffsetInNodeGroup);
     sizeColumn->scan(transaction, nodeGroupIdx, sizeColumnChunk.get(), startOffsetInNodeGroup,
         endOffsetInNodeGroup);
-    return {numOffsetsToRead, std::move(offsetColumnChunk), std::move(sizeColumnChunk)};
+    auto numValuesScan = offsetColumnChunk->getNumValues();
+    return {numValuesScan, std::move(offsetColumnChunk), std::move(sizeColumnChunk)};
 }
 
 void ListColumn::prepareCommitForChunk(Transaction* transaction, node_group_idx_t nodeGroupIdx,

diff --git a/src/storage/store/list_column_chunk.cpp b/src/storage/store/list_column_chunk.cpp
@@ -159,13 +159,17 @@ void ListColumnChunk::lookup(offset_t offsetInChunk, ValueVector& output,
     }
     auto startOffset = offsetInChunk == 0 ? 0 : getValue<offset_t>(offsetInChunk - 1);
     auto listSize = getListSize(offsetInChunk);
+    output.setValue<list_entry_t>(posInOutputVector, list_entry_t{startOffset, listSize});
     auto dataVector = ListVector::getDataVector(&output);
     auto currentListDataSize = ListVector::getDataVectorSize(&output);
     ListVector::resizeDataVector(&output, currentListDataSize + listSize);
     // TODO(Guodong): Should add `scan` interface and use `scan` here.
     for (auto i = 0u; i < listSize; i++) {
-        listDataColumnChunk->dataColumnChunk->lookup(startOffset + i, *dataVector, i);
+        listDataColumnChunk->dataColumnChunk->lookup(
+            startOffset + i, *dataVector, currentListDataSize + i);
     }
+    // reset offset
+    output.setValue<list_entry_t>(posInOutputVector, list_entry_t{currentListDataSize, listSize});
 }
 
 void ListColumnChunk::write(ColumnChunk* chunk, ColumnChunk* dstOffsets,

diff --git a/test/test_files/transaction/create_rel/list_disk_memory_scan.test b/test/test_files/transaction/create_rel/list_disk_memory_scan.test
@@ -0,0 +1,87 @@
+-GROUP TinySnbCreateNodeTest
+-DATASET CSV rel-update-tests
+
+--
+
+-DEFINE_STATEMENT_BLOCK INSERT_RELS_TO_SMALL_LIST [
+-STATEMENT MATCH (a:person),(b:person) WHERE a.ID=1 and b.ID=700 CREATE (a)-[e:knows{length: 45, tag: ['123', 'good']}]->(b)
+---- ok
+-STATEMENT MATCH (a:person),(b:person) WHERE a.ID=1 and b.ID=400 CREATE (a)-[e:knows{length: 34, place: 'short'}]->(b)
+---- ok
+-STATEMENT MATCH (a:person),(b:person) WHERE a.ID=1 and b.ID=600 CREATE (a)-[e:knows{}]->(b)
+---- ok
+]
+
+-CASE listNodeCreateDiskAndMemoryScan
+-STATEMENT CREATE NODE TABLE person2(ID INT64, workedHours INT64[], PRIMARY KEY (ID))
+---- ok
+-STATEMENT CREATE (:person2 {ID: 17, workedHours: [4,5,6]})
+---- ok
+-STATEMENT BEGIN TRANSACTION
+---- ok
+-STATEMENT CREATE (:person2 {ID: 12, workedHours: [2,3]})
+---- ok
+-STATEMENT CREATE (:person2 {ID: 14, workedHours: [2,5]})
+---- ok
+-STATEMENT MATCH (p:person2) RETURN p.workedHours
+---- 3
+[4,5,6]
+[2,3]
+[2,5]
+
+
+-CASE listRelInsertDiskAndMemoryScan
+-STATEMENT MATCH (a:person),(b:person) WHERE a.ID=1 and b.ID=300 CREATE (a)-[e:knows{length: 300, place: 'this is a long str', tag: ['123', 'good']}]->(b)
+---- ok
+-STATEMENT BEGIN TRANSACTION
+---- ok
+-INSERT_STATEMENT_BLOCK INSERT_RELS_TO_SMALL_LIST
+-STATEMENT MATCH (a:person)-[e:knows]->(b:person) where b.ID > 200 AND a.ID = 1 RETURN e.length, e.place, e.tag
+---- 4
+300|this is a long str|[123,good]
+34|short|
+45||[123,good]
+||
+
+
+-CASE listCreateEmptyDiskAndMemoryScan
+-STATEMENT CREATE NODE TABLE person2(ID INT64, workedHours INT64[], PRIMARY KEY (ID))
+---- ok
+-STATEMENT CREATE REL TABLE knows2(FROM person2 TO person2,  workedHours INT64[])
+---- ok
+-STATEMENT CREATE (:person2 {ID: 1,  workedHours: [4,5,6]})
+---- ok
+-STATEMENT BEGIN TRANSACTION
+---- ok
+-STATEMENT CREATE (:person2 {ID: 2, workedHours: [2,3]})
+---- ok
+-STATEMENT CREATE (:person2 {ID: 3, workedHours: [2,5]})
+---- ok
+-STATEMENT MATCH (p:person2) RETURN p.workedHours
+---- 3
+[4,5,6]
+[2,3]
+[2,5]
+-STATEMENT Commit
+---- ok
+-STATEMENT MATCH (p:person2) RETURN p.workedHours
+---- 3
+[4,5,6]
+[2,3]
+[2,5]
+-STATEMENT MATCH (a:person2),(b:person2) WHERE a.ID=1 and b.ID=2 CREATE (a)-[e:knows2{workedHours: [2,3,5,6,7,89]}]->(b)
+---- ok
+-STATEMENT BEGIN TRANSACTION
+---- ok
+-STATEMENT MATCH (a:person2),(b:person2) WHERE a.ID=2 and b.ID=3 CREATE (a)-[e:knows2{workedHours: [1,2,3]}]->(b)
+---- ok
+-STATEMENT MATCH (a:person2)-[e:knows2]->(b:person2) return e.workedHours
+---- 2
+[2,3,5,6,7,89]
+[1,2,3]
+-STATEMENT Commit
+---- ok
+-STATEMENT MATCH (a:person2)-[e:knows2]->(b:person2) return e.workedHours
+---- 2
+[2,3,5,6,7,89]
+[1,2,3]