topdog/supersonic
1
0
Fork 0
mirror of https://github.com/tencentmusic/supersonic.git synced 2025-12-14 22:25:19 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity

(fix)(headless)Fix schema corrector in that aliases should not be replaced.

Browse Source
This commit is contained in:
jerryjzhang
2025-03-12 16:31:43 +08:00
parent 0c8c2d4804
commit 9d13038599
4 changed files with 45 additions and 19 deletions
Download Patch File Download Diff File Expand all files Collapse all files

49
headless/chat/src/test/java/com/tencent/supersonic/headless/chat/corrector/AggCorrectorTest.java
View File

@@ -41,6 +41,27 @@ class AggCorrectorTest {
semanticParseInfo.getSqlInfo().getCorrectedS2SQL());
}
@Test
void testSchemaCorrector() {
SchemaCorrector corrector = new SchemaCorrector();
Long dataSetId = 1L;
ChatQueryContext chatQueryContext = buildQueryContext(dataSetId);
SemanticParseInfo semanticParseInfo = new SemanticParseInfo();
SchemaElement dataSet = new SchemaElement();
dataSet.setDataSetId(dataSetId);
semanticParseInfo.setDataSet(dataSet);
SqlInfo sqlInfo = new SqlInfo();
String sql = "WITH 总停留时长 AS (SELECT 用户, SUM(停留时长) AS _总停留时长_ FROM 超音数数据集 WHERE 用户 IN ('alice', 'lucy') AND 数据日期 >= '2025-03-01' AND 数据日期 <= '2025-03-12' GROUP BY 用户) SELECT 用户, _总停留时长_ FROM 总停留时长";
sqlInfo.setParsedS2SQL(sql);
sqlInfo.setCorrectedS2SQL(sql);
semanticParseInfo.setSqlInfo(sqlInfo);
corrector.correct(chatQueryContext, semanticParseInfo);
Assert.assertEquals(
"WITH 总停留时长 AS (SELECT 用户名, SUM(停留时长) AS _总停留时长_ FROM 超音数数据集 WHERE 用户名 IN ('alice', 'lucy') AND 数据日期 " +
">= '2025-03-01' AND 数据日期 <= '2025-03-12' GROUP BY 用户名) SELECT 用户名, _总停留时长_ FROM 总停留时长",
semanticParseInfo.getSqlInfo().getCorrectedS2SQL());
}
private ChatQueryContext buildQueryContext(Long dataSetId) {
ChatQueryContext chatQueryContext = new ChatQueryContext();
List<DataSetSchema> dataSetSchemaList = new ArrayList<>();
@@ -51,18 +72,30 @@ class AggCorrectorTest {
schemaElement.setDataSetId(dataSetId);
dataSetSchema.setDataSet(schemaElement);
Set<SchemaElement> dimensions = new HashSet<>();
SchemaElement element1 = new SchemaElement();
element1.setDataSetId(1L);
element1.setName("部门");
dimensions.add(element1);
dimensions.add(SchemaElement.builder()
.dataSetId(1L)
.name("部门")
.build());
dimensions.add(SchemaElement.builder()
.dataSetId(1L)
.name("用户名")
.build());
dataSetSchema.setDimensions(dimensions);
Set<SchemaElement> metrics = new HashSet<>();
SchemaElement metric1 = new SchemaElement();
metric1.setDataSetId(1L);
metric1.setName("访问次数");
metrics.add(metric1);
metrics.add(SchemaElement.builder()
.dataSetId(1L)
.name("访问次数")
.build());
metrics.add(SchemaElement.builder()
.dataSetId(1L)
.name("停留时长")
.build());
dataSetSchema.setMetrics(metrics);
dataSetSchemaList.add(dataSetSchema);