Closed williambrandler closed 5 months ago
issue verified. working on a fix.
proving more complex than originally thought. Requires a rewrite of some of the internals of the Piper
c'est la vie!
ok, thanks for the update.
How much effort do you think it will be?
thanks, this looks good
The new quarantine functionality in the pipe transformer will successfully run if there are corrupted records, however,
As an example, I created an input dataframe of 961 rows. The expected output dataframe count is 500 and the quarantine table is 461. What I see is 500 and 961. It seems like all the data is piped out into the quarantine table.
See screenshots below. Code to reproduce can be found here https://github.com/projectglow/glow/pull/461