Merge pull request #21 from pranavanba/main

Sync only the validated staging folders instead of entire bucket
Sage-Bionetworks · Feb 7, 2024 · b405f47 · b405f47
2 parents a7291fb + 547a60c
commit b405f47
Showing 1 changed file with 3 additions and 3 deletions.
diff --git a/staging_to_archive.R b/staging_to_archive.R
@@ -46,12 +46,12 @@ if (!is.null(synFindEntityId(validated_date, config::get("PARQUET_FOLDER_ARCHIVE
   sync_cmd <- glue::glue("aws s3 --profile service-catalog sync {STAGING_TO_ARCHIVE_DOWNLOAD_LOCATION} {base_s3_uri_archive}{validated_date}/ --exclude '*owner.txt*' --exclude '*archive*'")
   system(sync_cmd)
 
-  rm(sync_cmd, validated_date)
+  rm(sync_cmd)
 
-  # Sync entire bucket to local
+  # Sync new date dir in archive bucket to local
   unlink(STAGING_TO_ARCHIVE_DOWNLOAD_LOCATION, recursive = T, force = T)
   unlink(AWS_ARCHIVE_DOWNLOAD_LOCATION, recursive = T, force = T)
-  sync_cmd <- glue::glue('aws s3 --profile service-catalog sync {base_s3_uri_archive} {AWS_ARCHIVE_DOWNLOAD_LOCATION} --exclude "*owner.txt*" --exclude "*archive*"')
+  sync_cmd <- glue::glue('aws s3 --profile service-catalog sync {base_s3_uri_archive}{validated_date}/ {AWS_ARCHIVE_DOWNLOAD_LOCATION}/{validated_date}/ --exclude "*owner.txt*" --exclude "*archive*"')
   system(sync_cmd)
 
   # Modify cohort identifier in dir name