Compartilhar via

rxSparkCacheData {RevoScaleR}: Set the Cache Flag in Spark Compute Context

  • Artigo

Description

Use this function to set the cache flag to control whether data objects should be cached in Spark memory system, applicable for RxXdfData, RxHiveData, RxParquetData, and RxOrcData.

Usage

  rxSparkCacheData(data, cache = TRUE)

Arguments

data

a data source that can be RxXdfData, RxHiveData, RxParquetData, or RxOrcData. RxTextData is currently not supported.

cache

logical value controlling whether the data should be cached or not. If TRUEthe data will be cached in the Spark memory system after the first use for performance enhancement.

Details

Note that RxHiveData with saveAsTempTable=TRUE is always cached in memory regardless of the cache parameter setting in this function.

Value

data object with new cache flag.

See Also

rxSparkListData, rxSparkRemoveData, RxXdfData, RxHiveData, RxParquetData, RxOrcData.

Examples


 ## Not run:

 hiveData <- RxHiveData(table = "mytable")

 ## this does not work. The cache flag of the input hiveData is not changed, so it is still FALSE.
 rxSparkCacheData(hiveData)
 rxImport(hiveData) # data is not cached

 ## this works. The cache flag of hiveData is now set to TRUE.
 hiveData <- rxSparkCacheData(hiveData)  
 rxImport(hiveData) # data is now cached

 ## set cache value to FASLE
 hiveData <- rxSparkCacheData(hiveData, FALSE)
## End(Not run)