[ https://issues.apache.org/jira/browse/SPARK-24974?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
andrzej.stankev...@gmail.com updated SPARK-24974: ------------------------------------------------- Description: SharedInMemoryCache has all filestatus no matter whether you specify partition columns or not. It causes long load time for queries that use only couple partitions because Spark loads file's paths for files from all partitions. I partitioned files by *type* and i has directory structure like {code:java} report_date=2018-07-24/type=A/file_1 {code} I am trying to execute {code:java} val count = spark.read.parquet("/custom_path/report_date=2018-07-24").filter( "type == 'A'").count {code} In my query i need to load only files of type *A* and it is just a couple of files. But spark load all 19K of files from all partitions into SharedInMemoryCache which takes about 60 secs and only after that throws unused partitions. was: SharedInMemoryCache has all filestatus no matter whether you specify partition columns or not. It causes long load time for queries that use only couple partitions because Spark loads file's paths for files from all partitions. I partitioned files by *type* and i has directory structure like {code:java} report_date=2018-07-24/type=A/file_1 {code} I am trying to execute {code:java} val count = spark.read.parquet("/custom_path/report_date=2018-07-24").filter( "type == 'A'").count {code} In my query i need to load only files of type *A* and it is just a couple of files. But spark load all 19K of files from all partitions into SharedInMemoryCache which takes about 60 secs and only after that throws unused partitions. > Spark put all file's paths into SharedInMemoryCache even for unused > partitions. > ------------------------------------------------------------------------------- > > Key: SPARK-24974 > URL: https://issues.apache.org/jira/browse/SPARK-24974 > Project: Spark > Issue Type: Bug > Components: SQL > Affects Versions: 2.2.1 > Reporter: andrzej.stankev...@gmail.com > Priority: Major > > SharedInMemoryCache has all filestatus no matter whether you specify > partition columns or not. It causes long load time for queries that use only > couple partitions because Spark loads file's paths for files from all > partitions. > I partitioned files by *type* and i has directory structure like > {code:java} > report_date=2018-07-24/type=A/file_1 > {code} > > I am trying to execute > {code:java} > val count = spark.read.parquet("/custom_path/report_date=2018-07-24").filter( > "type == 'A'").count > {code} > > In my query i need to load only files of type *A* and it is just a couple of > files. But spark load all 19K of files from all partitions into > SharedInMemoryCache which takes about 60 secs and only after that throws > unused partitions. -- This message was sent by Atlassian JIRA (v7.6.3#76005) --------------------------------------------------------------------- To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org