These are chat archives for nextflow-io/nextflow

9th
May 2019
Evan Floden
@evanfloden
May 09 09:07

@stevekm

then tried to run my reference-file-download workflow in the same directory, and I could no longer use the cached results from the first workflow.

Was this the case even when using

nextflow run <pipeline> -resume <session id>

You can find the session id using nextflow log.

micans
@micans
May 09 09:11
@stevekm I discussed this with @pditommaso yesterday -- caching involves all inputs (values, files) and also the process definition and I think the script section (when committed to file). So it should work generally. That said, I prefer to separate any long tasks such as reference fetching and index building into a stand-alone pipeline, as it is easier to understand/maintain. But I know other people integrate it like you do.
@evanfloden congratulations man, the Netherlands are in mourning :-)
Evan Floden
@evanfloden
May 09 09:22
@micans unbelievable! I heard that only gooners use Snakemake #COYS
micans
@micans
May 09 09:26
hehe I get gooners and COYS (had to look that up though). Snakemake is just a new level of poetry!
Chelsea Sawyer
@csawye01
May 09 15:47
@pditommaso what is the main difference between Channel.create() and Channel.empty()? I noticed if I want to optionally put several channels into an empty channel when using Channel.create() my pipeline will hang (waiting for more input) where as if I use Channel.empty() and put several channels into it, my pipeline will complete.
micans
@micans
May 09 16:56
@csawye01 I wonder what you are doing/aiming for. What do you do when you put several channels into an empty channel when using Channel.create() ?