Re: [Treebase-devel] Some data problems

[William P. <wil...@ya...>]

On Jan 29, 2010, at 12:01 PM, Hilmar Lapp wrote:

> I guess that need not necessarily be resolved before the release, but it seems to me that there is a major data cleanup effort necessary at some point, combined with enforcing all possible constraints in the database.

I agree that this cleanup can be done after release -- there are so few studies in question that this can be done by hand. 

The main thing to check before release is that there are no current bugs that are creating this mis-wiring. Once we have a working sequence and the boolean-bug is fixed, I'll upload a bunch of test datasets and then rerun these queries to check that no new mis-wiring is happening. 

One thing I noticed is that when I upload a tree to a submission, and then delete the tree, the taxon labels are not deleted.  Over time, this will created a clutter of taxon labels that don't belong to any tree or matrix. Should we call this a bug? If so, fixing it will require some effort: upon deletion of a tree or a matrix, the software must know to delete any related taxon labels unless they are being used by another matrix, rowsegment, or tree that still exists in the study. 

bp