Fix df.Show() issue

Author: arjunsk

README.md

@@ -28,6 +28,8 @@ import (
 func TestParquetFile(t *testing.T) {
 	ctx := engine.NewContext()
 	df, err := ctx.Parquet("../../test/data/c1_c2_c3_int64.parquet", nil)
+	
+	_ = df.Show()
 	/*
 	+-----+-----+-----+
 	| C1  | C2  | C3  |

cmd/simple/engine_test.go

@@ -15,6 +15,8 @@ func TestParquetFile(t *testing.T) {
 		t.Error(err)
 	}
 
+	_ = df.Show()
+
 	df = df.
 		Filter(logicalplan.Eq(
 			logicalplan.ColumnExpr{Name: "c1"},

pkg/b_dataframe/dataframe.go

@@ -63,12 +63,26 @@ func (df *DataFrame) Aggregate(groupBy []logicalplan.Expr, aggExpr []logicalplan
 }
 
 func (df *DataFrame) collect(ctx context.Context, callback datasource.Callback) error {
-	df.planBuilder.Output(callback)
+	// create a copy of the plan builder and add the output operator
+	// NOTE: This is a hack to add Output operator to the PhysicalPlan.
+	builder := df.planBuilder.Clone().Output(callback)
 
-	physicalPlan, err := df.PhysicalPlan()
+	// build the logical plan
+	plan, err := builder.Build()
 	if err != nil {
 		return err
 	}
+
+	// optimize the logical plan
+	plan = df.ruleBasedOptimizer.Optimize(plan)
+
+	// create the physical plan
+	physicalPlan, err := df.sessionState.CreatePhysicalPlan(plan)
+	if err != nil {
+		return err
+	}
+
+	// execute the physical plan
 	return physicalPlan.Execute(df.TaskContext(), callback)
 }
 

pkg/c_logical_plan/f_logical_plan_builder.go

@@ -44,6 +44,14 @@ func (b *Builder) Build() (LogicalPlan, error) {
 	return b.plan, nil
 }
 
+func (b *Builder) Clone() *Builder {
+	var copiedPlan LogicalPlan
+	if b.plan != nil {
+		copiedPlan = b.plan
+	}
+	return &Builder{plan: copiedPlan}
+}
+
 func Validate(plan LogicalPlan) error {
 	return nil
 }