Re: [Treebase-devel] Fwd: [[treebase-help] #7630] TreeBASE Submission S10318

[youjun g. <you...@ya...>]

>
> In theory making it 100% Hibernate should solve the problem, and should
> also make it consistent with the Hibernate ORM tier for data access. The
> more straight SQL we have in there, the more maintenance problems or risks
> we have down the road. Why does a pure Hibernate solution not work here?
> Hibernate is perfectly capable of orchestrating cascading deletes.


>
>        -hilmar
>
>
> Based on the information I get from Bill, the historic reason why Jin Ruan
bring in SQL here is because the performance issue of hibernate.

I think the sql created by hibernate is not as efficient when a task
involved too many tables and too many cascade deletions.

To delete a matrix we need to delete all related record from

study,
submission,
matrix,
sub_matrix
matrixrow,
rowsegment
matrixcolumn,
MATRIXCOLUMN_ITEMDEFINITION
matrixelemnt,
STATEMODIFIER
ITEMVALUE
Compound_element
taxonlabelset
taxonlabelset_taxonlabel
taxonlabel


and also need to check a group of tables to see make sure a taxonlabel  is
not refered by some other matrices or trees.

While listing of the steps above I think our database schema may need a
optimizing some how.

Youjun