We've moved!
You can find our new documentation site and support forum for posting questions here.

Snapshotting / Timestamping of data attributes

ericco92ericco92 Cambridge, UKMember ✭✭

I'm finding that we often end up rerunning analyses that update sample/pair/pairset attributes. We also do local updates to input files and push those to workspace attributes. However, these don't get timestamped, making it impossible to tell which workflow (or workflow snapshot) actually generated these files. It can quickly become hard to know which version of a workflow or upload generated a file (without picking through the Monitor tab).

A creation timestamp for data / workspace attributes, visible on the table browser, would go a long way to sorting this out.

3 votes

Active · Last Updated


  • RobinKRobinK Member

    Agree! A way to link to which run (and snapshot) an attribute was generated by would be very helpful for provenance tracking.

  • ericco92ericco92 Cambridge, UKMember ✭✭

    I ran into another issue with this today, where a workflow was run using older data. I had to crawl through the FireCloud UI (and eventually three different google buckets) to establish that one of my data attributes (which are easily moved between workspaces, but without timestamps or provenance) was from an older version of a prerequisite program. This caused a whole chain of analyses to be invalidated. Do the logs have this info, maybe as an md5sum or something?

Sign In or Register to comment.