[ https://issues.apache.org/jira/browse/SPARK-15174?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15273805#comment-15273805 ]
hujiayin edited comment on SPARK-15174 at 5/12/16 4:09 AM: ----------------------------------------------------------- I could fix it. was (Author: hujiayin): @Ian Hellstrom, I think the issue is caused by "if (groupingExpressions.isEmpty) { Statistics(sizeInBytes = 1) }" in basicLogicOperators.scala. I could fix it for you. > DataFrame does not have correct number of rows after dropDuplicates > ------------------------------------------------------------------- > > Key: SPARK-15174 > URL: https://issues.apache.org/jira/browse/SPARK-15174 > Project: Spark > Issue Type: Bug > Components: SQL > Affects Versions: 1.4.1 > Reporter: Ian Hellstrom > > If you read an empty file/folder with the {{SQLContext.read()}} function and > call {{DataFrame.dropDuplicates()}}, the number of rows is incorrect. > {code} > val input = "hdfs:///some/empty/directory" > val df1 = sqlContext.read.json(input) > val df2 = sqlContext.read.json(input).dropDuplicates > df1.count == 0 // true > df1.rdd.isEmpty // true > df2.count == 0 // false: it's actually reported as 1 > df2.rdd.isEmpty // false > {code} -- This message was sent by Atlassian JIRA (v6.3.4#6332) --------------------------------------------------------------------- To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org