Hi there 👋 I am trying to set up an S3 sensor and I'm definitely doing something wrong but I can't spot it. If a key enters the bucket that isn't just a new folder and isn't in the

processed/

folder, then I want to run

my_job

and give it the name of a SQL script to run. The idea is that we may want to run different scripts depending on the file, but I haven't set that part up yet. When a file is processed, an op copies that file into a

processed/

folder in the same S3 bucket, thus why I want to yield a SkipReason for those files as well. When I turn on this sensor, I'm getting the following error:

dagster.check.CheckError: Failure condition: Expected a single SkipReason or one or more RunRequests: received both RunRequest and SkipReason

I'll add the sensor definition in thread (slightly altered to hide my bucket name)

within your loop, it looks like it is possible for one key to match the first condition, which yield a

SkipReason

, and the next key to pass all filters and yield a

RunRequest

.

Okay I see what you're saying. Hmm. I guess what I was trying to account for is when multiple keys get added in the S3 bucket at a time but I only want to run the job for some of the keys and not others. In that case, should I instead just not yield a SkipReason for the keys I don't want, but still yield RunRequests for the keys I do want?

yeah, you'll need to just rely on logging those exempted keys if you're going to yield at least one RunRequest, it looks like

That did the trick. Thanks a lot for the help!