[jira] [Created] (CALCITE-7415) CalciteCatalogReader.lookupOperatorOverloads keeps original function identifier casing instead of resolved schema-path casing

[zzwqqq (Jira)]

zzwqqq created CALCITE-7415:
-------------------------------

             Summary: CalciteCatalogReader.lookupOperatorOverloads keeps 
original function identifier casing instead of resolved schema-path casing
                 Key: CALCITE-7415
                 URL: https://issues.apache.org/jira/browse/CALCITE-7415
             Project: Calcite
          Issue Type: Bug
          Components: core
            Reporter: zzwqqq


CalciteCatalogReader#lookupOperatorOverloads resolves the function from schema, 
but the operator name identifier may still keep the original input casing 
(catalog/schema/function).
As a result, SqlUserDefinedFunction / RelToSql may unparse qualified UDF names 
with incorrect case, which can break SQL execution on case-sensitive dialects 
or environments.

Steps to reproduce：
{code:java}
import org.apache.calcite.config.Lex;
import org.apache.calcite.rel.RelNode;
import org.apache.calcite.rel.rel2sql.RelToSqlConverter;
import org.apache.calcite.rel.type.RelDataType;
import org.apache.calcite.rel.type.RelDataTypeFactory;
import org.apache.calcite.schema.SchemaPlus;
import org.apache.calcite.schema.impl.AbstractSchema;
import org.apache.calcite.schema.impl.AbstractTable;
import org.apache.calcite.schema.impl.ScalarFunctionImpl;
import org.apache.calcite.sql.SqlNode;
import org.apache.calcite.sql.dialect.OracleSqlDialect;
import org.apache.calcite.sql.parser.SqlParser;
import org.apache.calcite.sql.type.SqlTypeName;
import org.apache.calcite.tools.FrameworkConfig;
import org.apache.calcite.tools.Frameworks;
import org.apache.calcite.tools.Planner;

public class FunctionIdentifierCaseDemo {
  public static class UdfLib {
    public static int MyFunc(int x) {
      return x;
    }
  }

  public static void main(String[] args) throws Exception {
    SchemaPlus root = Frameworks.createRootSchema(true);
    SchemaPlus catalog = root.add("MyCatalog", new AbstractSchema());
    SchemaPlus schema = catalog.add("MySchema", new AbstractSchema());

    schema.add("Emp", new AbstractTable() {
      @Override public RelDataType getRowType(RelDataTypeFactory typeFactory) {
        return typeFactory.builder()
            .add("EMPNO", SqlTypeName.INTEGER)
            .build();
      }
    });
    schema.add("MyFunc", ScalarFunctionImpl.create(UdfLib.class, "MyFunc"));

    FrameworkConfig config = Frameworks.newConfigBuilder()
        .defaultSchema(root)
        .parserConfig(SqlParser.config()
            .withLex(Lex.ORACLE)
            .withCaseSensitive(false))
        .build();

    String query = "SELECT MYCATALOG.MYSCHEMA.MYFUNC(empno) "
        + "FROM MYCATALOG.MYSCHEMA.EMP";

    try (Planner planner = Frameworks.getPlanner(config)) {
      SqlNode parsed = planner.parse(query);
      SqlNode validated = planner.validate(parsed);
      RelNode rel = planner.rel(validated).project();

      SqlNode outNode = new RelToSqlConverter(OracleSqlDialect.DEFAULT)
          .visitRoot(rel)
          .asStatement();
      String out = outNode.toSqlString(OracleSqlDialect.DEFAULT).getSql();

      System.out.println("Input SQL:");
      System.out.println(query);
      System.out.println("\nRelToSql (Oracle):");
      System.out.println(out);
    }
  }
} {code}
expected generated sql：
{code:java}
RelToSql (Oracle):
SELECT "MyCatalog"."MySchema"."MyFunc"("EMPNO")
FROM "MyCatalog"."MySchema"."Emp" {code}
but actual is：
{code:java}
RelToSql (Oracle):
SELECT "MYCATALOG"."MYSCHEMA"."MYFUNC"("EMPNO")
FROM "MyCatalog"."MySchema"."Emp" {code}



--
This message was sent by Atlassian Jira
(v8.20.10#820010)