From: William P. <wil...@ya...> - 2010-10-22 14:48:13
|
On Oct 22, 2010, at 9:18 AM, Hilmar Lapp wrote: > That's a known issue. We delete those "by hand" mostly, where by hand is running a script. > > -hilmar Yeah... for added nuance: submissions can be deleted if the trees and matrices inside them are deleted first. Course, in your case, the added problem is that the 720 x 6309 matrix and tree are themselves reticent to delete. I did manage to get this matrix deleted in Martin Wu's account: even though there is a proxy time-out, the program seems to still grind on, eventually deleting the data. A 720 x 6309 matrix has about 5 million records to delete, and it looks like Hibernate's approach is inefficient. The tree, however, still refuses to delete, which I find weird because 720 taxa is not a big tree. But, as Hilmar mentioned, we have a perl script to delete tree blocks -- this has successfully deleted tree blocks with 5,000 trees in them of 150 taxa each, so deleting a single tree of 720 taxa shouldn't be a problem. bp |