lazy val withCachedData: LogicalPlan = {
    assertAnalyzed()
    assertSupported()
    sparkSession.sharedState.cacheManager.useCachedData(analyzed)
  }

/** Replaces segments of the given logical plan with cached versions where possible. */
def useCachedData(plan: LogicalPlan): LogicalPlan = {
  val newPlan = plan transformDown {
    // Do not lookup the cache by hint node. Hint node is special, we should ignore it when
    // canonicalizing plans, so that plans which are same except hint can hit the same cache.
    // However, we also want to keep the hint info after cache lookup. Here we skip the hint
    // node, so that the returned caching plan won‘t replace the hint node and drop the hint info
    // from the original plan.
    case hint: ResolvedHint => hint

    case currentFragment =>
      lookupCachedData(currentFragment)
        .map(_.cachedRepresentation.withOutput(currentFragment.output))
        .getOrElse(currentFragment)
  }

  newPlan transformAllExpressions {
    case s: SubqueryExpression => s.withNewPlan(useCachedData(s.plan))
  }

/** Optionally returns cached data for the given [[LogicalPlan]]. */
  def lookupCachedData(plan: LogicalPlan): Option[CachedData] = readLock {
    cachedData.asScala.find(cd => plan.sameResult(cd.plan))
  }

@transient
private val cachedData = new java.util.LinkedList[CachedData]

/** Holds a cached logical plan and its data */
case class CachedData(plan: LogicalPlan, cachedRepresentation: InMemoryRelation)

/**
   * Caches the data produced by the logical representation of the given [[Dataset]].
   * Unlike `RDD.cache()`, the default storage level is set to be `MEMORY_AND_DISK` because
   * recomputing the in-memory columnar representation of the underlying table is expensive.
   */
  def cacheQuery(
      query: Dataset[_],
      tableName: Option[String] = None,
      storageLevel: StorageLevel = MEMORY_AND_DISK): Unit = writeLock {
    val planToCache = query.logicalPlan
    if (lookupCachedData(planToCache).nonEmpty) {
      logWarning("Asked to cache already cached data.")
    } else {
      val sparkSession = query.sparkSession
      val inMemoryRelation = InMemoryRelation(
        sparkSession.sessionState.conf.useCompression,
        sparkSession.sessionState.conf.columnBatchSize, storageLevel,
        sparkSession.sessionState.executePlan(AnalysisBarrier(planToCache)).executedPlan,
        tableName,
        planToCache.stats)
      cachedData.add(CachedData(planToCache, inMemoryRelation))
    }
  }

def persist(): this.type = {
  sparkSession.sharedState.cacheManager.cacheQuery(this)
  this
}