add setJobAutoDelete function

Author: zfengms

CHANGELOG.md

@@ -1,17 +1,15 @@
 # Change Log
-## [0.6.2] 2017-11-17
+
+## [0.6.1] 2017-12-05
 ### Added
+- Support for users to use programmatically generated credentials and cluster config
 - Support for users to delete job and terminate job
 ### Changed
-- [BREAKING CHANGE] when wait = TRUE, both job and job results are deleted at the end of the run
+- [BREAKING CHANGE] when wait = TRUE, both job and job results are deleted at the end of the run, set jobAutoComplete to FALSE to keep them
 - Add retry to get job result
 - Add errorHandling and wait option to job metadata
 - Save job metadata to job result storage blob
 
-## [0.6.1] 2017-11-13
-### Added
-- Support for users to use programmatically generated credentials and cluster config
-
 ## [0.6.0] 2017-11-03
 ### Added
 - Support for users to run custom versions of R via Docker containers

NAMESPACE

@@ -20,6 +20,7 @@ export(resizeCluster)
 export(setChunkSize)
 export(setCredentials)
 export(setHttpTraffic)
+export(setJobAutoDelete)
 export(setReduce)
 export(setVerbose)
 export(stopCluster)

R/doAzureParallel.R

@@ -73,6 +73,20 @@ setChunkSize <- function(value = 1) {
   assign("chunkSize", value, envir = .doAzureBatchGlobals)
 }
 
+#' Specify whether to delete job/jobresult after asychronous job is completed.
+#'
+#' @param value boolean of TRUE or FALSE
+#'
+#' @examples
+#' setJobAutoDelete(FALSE)
+#' @export
+setJobAutoDelete <- function(value = TRUE) {
+  if (!is.logical(value))
+    stop("setJobAutoDelete requires a boolean argument")
+
+  assign("jobAutoDelete", value, envir = .doAzureBatchGlobals)
+}
+
 #' Apply reduce function on a group of iterations of the foreach loop together per task.
 #'
 #' @param fun The number of iterations to group
@@ -232,10 +246,16 @@ setHttpTraffic <- function(value = FALSE) {
     wait <- obj$options$azure$wait
   }
 
-  # by default, delete both job and job result after job is completed (both synchronous and asynchronous)
-  deleteJob <- TRUE
-  if (!is.null(obj$options$azure$deleteJob)) {
-    wait <- obj$options$azure$deleteJob
+  # by default, delete both job and job result after synchronous job is completed
+  jobAutoDelete <- TRUE
+
+  if (exists("jobAutoDelete", envir = .doAzureBatchGlobals)) {
+    jobAutoDelete <- get("jobAutoDelete", envir = .doAzureBatchGlobals)
+  }
+
+  if (!is.null(obj$options$azure$jobAutoDelete) &&
+      is.logical(obj$options$azure$jobAutoDelete)) {
+    jobAutoDelete <- obj$options$azure$jobAutoDelete
   }
 
   inputs <- FALSE
@@ -288,6 +308,10 @@ setHttpTraffic <- function(value = FALSE) {
 
   chunkSize <- 1
 
+  if (exists("chunkSize", envir = .doAzureBatchGlobals)) {
+    chunkSize <- get("chunkSize", envir = .doAzureBatchGlobals)
+  }
+
   if (!is.null(obj$options$azure$chunkSize)) {
     chunkSize <- obj$options$azure$chunkSize
   }
@@ -296,10 +320,6 @@ setHttpTraffic <- function(value = FALSE) {
     chunkSize <- obj$options$azure$chunksize
   }
 
-  if (exists("chunkSize", envir = .doAzureBatchGlobals)) {
-    chunkSize <- get("chunkSize", envir = .doAzureBatchGlobals)
-  }
-
   chunkSizeKeyValuePair <-
     list(name = "chunkSize", value = as.character(chunkSize))
 
@@ -565,7 +585,7 @@ setHttpTraffic <- function(value = FALSE) {
 
           numberOfFailedTasks <- sum(unlist(failTasks))
 
-          if (numberOfFailedTasks > 0 && deleteJob == FALSE) {
+          if (numberOfFailedTasks > 0 && jobAutoDelete == FALSE) {
             .createErrorViewerPane(id, failTasks)
           }
 
@@ -587,7 +607,7 @@ setHttpTraffic <- function(value = FALSE) {
               fill = TRUE)
 
           # delete job from batch service and job result from storage blob
-          if (deleteJob) {
+          if (jobAutoDelete) {
             deleteJob(id)
           }
 

R/jobUtilities.R

@@ -252,19 +252,14 @@ getJobResult <- function(jobId) {
 #' Delete a job
 #'
 #' @param jobId A job id
-#' @param deleteResult TRUE to delete job result in storage blob
-#' container, FALSE to keep the result in storage blob container.
 #'
 #' @examples
 #' \dontrun{
 #' deleteJob("job-001")
-#' deleteJob("job-001", deleteResult = FALSE)
 #' }
 #' @export
-deleteJob <- function(jobId, deleteResult = TRUE) {
-  if (deleteResult == TRUE) {
-    deleteStorageContainer(jobId)
-  }
+deleteJob <- function(jobId) {
+  deleteStorageContainer(jobId)
 
   response <- rAzureBatch::deleteJob(jobId, content = "response")
 

docs/31-long-running-job.md

@@ -43,6 +43,8 @@ getJob returns job metadata, such as chunk size, whether cloud combine is enable
 		succeeded: 0
 		failed: 5
 	total: 6
+
+  job state: completed
 ```
 
 ## Get job list
@@ -81,14 +83,14 @@ Once job is completed successfully, you can call getJobResult to retrieve the jo
 
 ### Clean up
 
-Once you get the job result, you can delete the job.
+Once you get the job result, you can delete the job and its result.
 ```R
   deleteJob(jobId)
 ```
 
-Please note deleteJob will delete the job at batch service, by default, it also deletes the storage container holding the job result. If you want to keep the job result around, you can set deleteResult parameter to FALSE
+Please note deleteJob will delete the job at batch service and the storage container holding the job result.
 ```R
-  deleteJob(jobId, deleteResult = FALSE)
+  deleteJob(jobId)
 ```
 
 A [working sample](../samples/long_running_job/long_running_job.R) can be found in the samples directory.

man/deleteJob.Rd

@@ -65,4 +65,4 @@ jobResult <- getJobResult(jobId)
 doAzureParallel::stopCluster(cluster)
 
 # delete the job
-rAzureBatch::deleteJob(jobId)
+deleteJob(jobId)

man/setJobAutoDelete.Rd

@@ -52,6 +52,6 @@ test_that("Long Running Job Test", {
   testthat::expect_equal(jobResult,
                          list(2, 2, 2, 2))
 
-  # delete the job
+  # delete the job and its result
   deleteJob(jobId)
 })