Hey all! is there an easy way to view the intermed...
# ask-community
Hey all! is there an easy way to view the intermediate output between steps without yielding an asset or writing to a csv (for local dev)? For example, if I wanted to read a dataframe, drop the cols, process the df, etc. (See screenshot) for example, if I wanted to view the data coming in as a dataframe could I do it from the UI (to keep everything in one place)? I did a quick look through Slack/Google/Docs and couldn't seem to find an easy way to do this :/
If "without yielding an asset" is not a strong requirement, you can try to use pandas' to_markdown and yield
with `EventMetadataEntry.md` event (you don't have to actually persist dataframe anywhere)
hi @Mathew Lee -- another option to the above suggestion would be to attach that same metadata entry to the output. So instead of just returning the output dataframe, you could
yield Output(df, metadata={"data": EventMetadata.md(<http://df.to|df.to>_markdown())})
. This metadata will show up in the event log attached to the output. Docs here: https://docs.dagster.io/concepts/ops-jobs-graphs/op-events#attaching-metadata-to-outputs