[jira] [Updated] (SPARK-29272) dataframe.write.format("libsvm").save() take too much time

["张焕明 (Jira)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/SPARK-29272?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

张焕明 updated SPARK-29272:
------------------------
    Summary: dataframe.write.format("libsvm").save() take too much time  (was: dataframe.write.format("libsvm").save() 保存时间太长)

> dataframe.write.format("libsvm").save() take too much time
> ----------------------------------------------------------
>
>                 Key: SPARK-29272
>                 URL: https://issues.apache.org/jira/browse/SPARK-29272
>             Project: Spark
>          Issue Type: Question
>          Components: ML
>    Affects Versions: 2.2.0
>            Reporter: 张焕明
>            Priority: Major
>
> 我有一个pyspark的dataframe，小数据测试，只有3M，当调用
> MLUtils.convertVectorColumnsToML(dataframe).write.format("libsvm").save('path'), mode='overwrite')时，花费10秒，而当我用filter函数只筛选出10条数据再保存
> temp_df = dataframe.filter(train_df['__index'].between(int(0,10))，仍然需要8s的时间开销，为什么那么慢啊，我需要怎么改进吗



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

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