spark-reviews mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From srinathshankar <...@git.apache.org>
Subject [GitHub] spark pull request #15090: [SPARK-17073] [SQL] generate column-level statist...
Date Wed, 21 Sep 2016 20:12:44 GMT
Github user srinathshankar commented on a diff in the pull request:

    https://github.com/apache/spark/pull/15090#discussion_r79893014
  
    --- Diff: sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/plans/logical/Statistics.scala
---
    @@ -32,19 +34,74 @@ package org.apache.spark.sql.catalyst.plans.logical
      * @param sizeInBytes Physical size in bytes. For leaf operators this defaults to 1,
otherwise it
      *                    defaults to the product of children's `sizeInBytes`.
      * @param rowCount Estimated number of rows.
    + * @param colStats Column-level statistics.
      * @param isBroadcastable If true, output is small enough to be used in a broadcast join.
      */
     case class Statistics(
         sizeInBytes: BigInt,
         rowCount: Option[BigInt] = None,
    +    colStats: Map[String, ColumnStats] = Map.empty,
         isBroadcastable: Boolean = false) {
    +
       override def toString: String = "Statistics(" + simpleString + ")"
     
       /** Readable string representation for the Statistics. */
       def simpleString: String = {
         Seq(s"sizeInBytes=$sizeInBytes",
           if (rowCount.isDefined) s"rowCount=${rowCount.get}" else "",
    +      if (colStats.nonEmpty) s"colStats=$colStats" else "",
           s"isBroadcastable=$isBroadcastable"
    -    ).filter(_.nonEmpty).mkString("", ", ", "")
    +    ).filter(_.nonEmpty).mkString(", ")
    +  }
    +}
    +
    +/**
    + * Statistics for a column.
    + * @param ndv Number of distinct values of the column.
    + */
    +case class ColumnStats(
    +    dataType: DataType,
    +    numNulls: Long,
    +    max: Option[Any] = None,
    +    min: Option[Any] = None,
    +    ndv: Option[Long] = None,
    +    avgColLen: Option[Double] = None,
    +    maxColLen: Option[Long] = None,
    +    numTrues: Option[Long] = None,
    +    numFalses: Option[Long] = None) {
    +
    +  override def toString: String = "ColumnStats(" + simpleString + ")"
    +
    +  def simpleString: String = {
    +    Seq(s"numNulls=$numNulls",
    +      if (max.isDefined) s"max=${max.get}" else "",
    +      if (min.isDefined) s"min=${min.get}" else "",
    +      if (ndv.isDefined) s"ndv=${ndv.get}" else "",
    +      if (avgColLen.isDefined) s"avgColLen=${avgColLen.get}" else "",
    +      if (maxColLen.isDefined) s"maxColLen=${maxColLen.get}" else "",
    +      if (numTrues.isDefined) s"numTrues=${numTrues.get}" else "",
    +      if (numFalses.isDefined) s"numFalses=${numFalses.get}" else ""
    +    ).filter(_.nonEmpty).mkString(", ")
    +  }
    +}
    +
    +object ColumnStats {
    +  def fromString(str: String, dataType: DataType): ColumnStats = {
    --- End diff --
    
    Suggestion: Based on fromString, It looks like simpleString above is practically used
as a serializer to stuff stats in the catalog ? If so, could you change the name to catalogRepresentation
or something like that so it's clear.


---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastructure@apache.org or file a JIRA ticket
with INFRA.
---

---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org


Mime
View raw message