You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
When running cmflib/commands/artifact/pull.py , we are fetching all executions of a stage for a pipeline. Is there a guarantee that the executions returned will be in time order. Perhaps we should introduce an option to return a time ordered list. For example:
get_all_executions_in_stage(stage, ordered=True)
This becomes important if there are multiple executions (which have produced a given artifact) and the latest one needs to be pulled/returned. At this point all are returned, but which is the one that will be written to the remote filesystem? If there are multiple file hashes corresponding to a named file, its a matter of chance, what is pulled last. (Without time-ordering)
From pull.py run()
executions = query.get_all_executions_in_stage(stage)
# check if stage has executions
if len(executions) > 0:
# converting it to dictionary
dict_executions = executions.to_dict("dict")
# append id's of executions inside identifiers
for id in dict_executions["id"].values():
identifiers.append(id)
The text was updated successfully, but these errors were encountered:
When running cmflib/commands/artifact/pull.py , we are fetching all executions of a stage for a pipeline. Is there a guarantee that the executions returned will be in time order. Perhaps we should introduce an option to return a time ordered list. For example:
get_all_executions_in_stage(stage, ordered=True)
This becomes important if there are multiple executions (which have produced a given artifact) and the latest one needs to be pulled/returned. At this point all are returned, but which is the one that will be written to the remote filesystem? If there are multiple file hashes corresponding to a named file, its a matter of chance, what is pulled last. (Without time-ordering)
From pull.py run()
The text was updated successfully, but these errors were encountered: