Have you tried to use data libraries for this? There's an import mechanism there that'll allow you to simply link to the file on disk without copy/upload. I believe the "example_watch_folder.py" sample script (in the distribution) does just this via the API, if you want an example.

On Mon, Sep 23, 2013 at 5:15 PM, Ted Goldstein <ted@soe.ucsc.edu> wrote:

I want to frequently import many tens of thousands of datasets. The files are on the same sever as Galaxy. But the upload based mechanism is really really slow. It takes hours to load this many files, yet the data is not moving at all!

What is the best strategy to go about making a faster bulk import? I can imagine a tight loop that is

My datatypes are limited.

foreach folder
new LibraryFolder

foreach file in each directory
new LibraryDataset
new LibraryDatasetDatasetAssociation

flush once at the end.

Thoughts?

___________________________________________________________
Please keep all replies on the list by using "reply all"
in your mail client. To manage your subscriptions to this
and other Galaxy lists, please use the interface at:
http://lists.bx.psu.edu/

To search Galaxy mailing lists use the unified search at:
http://galaxyproject.org/search/mailinglists/