[SPARK-9973] [SQL] Correct in-memory columnar buffer size

The `initialSize` argument of `ColumnBuilder.initialize()` should be the
number of rows rather than bytes.  However `InMemoryColumnarTableScan`
passes in a byte size, which makes Spark SQL allocate more memory than
necessary when building in-memory columnar buffers.

Author: Kun Xu <[email protected]>

Closes apache#8189 from viper-kun/errorSize.

Author: viper-kun

sql/core/src/main/scala/org/apache/spark/sql/columnar/InMemoryColumnarTableScan.scala

@@ -121,8 +121,7 @@ private[sql] case class InMemoryRelation(
         def next(): CachedBatch = {
           val columnBuilders = output.map { attribute =>
             val columnType = ColumnType(attribute.dataType)
-            val initialBufferSize = columnType.defaultSize * batchSize
-            ColumnBuilder(attribute.dataType, initialBufferSize, attribute.name, useCompression)
+            ColumnBuilder(attribute.dataType, batchSize, attribute.name, useCompression)
           }.toArray
 
           var rowCount = 0