[CALCITE-3319] AssertionError for ReduceDecimalsRule

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 1.22.0
• Labels: pull-request-available
• Assignee: Unassigned
• Reporter: Wang Yanlin
• Created: 03/Sep/19 13:27
• Updated: 06/Mar/20 04:33
• Resolved: 12/Sep/19 16:41

Description

When trying to use ReduceDecimalsRule, I got

java.lang.AssertionError
	at org.apache.calcite.rel.rules.ReduceDecimalsRule$DecimalShuttle.visitCall(ReduceDecimalsRule.java:159)
	at org.apache.calcite.rel.rules.ReduceDecimalsRule$DecimalShuttle.visitCall(ReduceDecimalsRule.java:124)
	at org.apache.calcite.rex.RexCall.accept(RexCall.java:191)
	at org.apache.calcite.rex.RexProgramBuilder.add(RexProgramBuilder.java:653)
	at org.apache.calcite.rex.RexProgramBuilder.create(RexProgramBuilder.java:601)
	at org.apache.calcite.rel.rules.ReduceDecimalsRule.onMatch(ReduceDecimalsRule.java:103)
	at org.apache.calcite.plan.AbstractRelOptPlanner.fireRule(AbstractRelOptPlanner.java:319)
	at org.apache.calcite.plan.hep.HepPlanner.applyRule(HepPlanner.java:560)
	at org.apache.calcite.plan.hep.HepPlanner.applyRules(HepPlanner.java:419)
	at org.apache.calcite.plan.hep.HepPlanner.executeInstruction(HepPlanner.java:256)
	at org.apache.calcite.plan.hep.HepInstruction$RuleInstance.execute(HepInstruction.java:127)
	at org.apache.calcite.plan.hep.HepPlanner.executeProgram(HepPlanner.java:215)
	at org.apache.calcite.plan.hep.HepPlanner.findBestExp(HepPlanner.java:202)

I read the code, and found this.

List<RexNode> newOperands = apply(call.getOperands());
      if (true) {
        // FIXME: Operands are now immutable. Create a new call with
        //   new operands?
        throw new AssertionError();
      }

After remove this logic, the rel below

LogicalCalc(expr#0..7=[{inputs}], expr#8=[1.01E1:DOUBLE], expr#9=[15], expr#10=[+($t8, $t9)], expr#11=[>($t5, $t10)], proj#0..7=[{exprs}], $condition=[$t11])
  LogicalTableScan(table=[[scott, EMP]])

can be translated into

LogicalCalc(expr#0..7=[{inputs}], expr#8=[Reinterpret($t5)], expr#9=[CAST($t8):DOUBLE], expr#10=[1E2:DOUBLE], expr#11=[/INT($t9, $t10)], expr#12=[1.01E1:DOUBLE], expr#13=[15], expr#14=[+($t12, $t13)], expr#15=[>($t11, $t14)], proj#0..7=[{exprs}], $condition=[$t15])
  LogicalTableScan(table=[[scott, EMP]])

So is this rule not ready for use now, or we should just remove this logic?

Issue Links

• links to: GitHub Pull Request #1442