Use PreparedStatement instead of looping Table.insert

[bbimber]

@labkey-martyp, this is tangential to #730. Now that the kinship process is actually working, I forgot how long the insert takes. Matt keyed me in to using PreparedStatement instead of looping over Table.insert(). This should be a minor change that improves insert perf.

[labkey-martyp]

This is great. Thanks.

[bbimber]

hi @labkey-martyp, I just added one additional small change. this commit allows the job to monitor for being cancelled, and respect that. you already approved, but given this is a change I thought I'd ping you on it.

Also, I added a sort on the TSV produced by the kinship R script. This should have no functional change with import now, but it occurred to me that given how rarely these data change, it is probably unnecessary to constantly re-import everything. If PreparedStatement isnt enough of a speed boost, the other option would be to read two streams: one from the TSV and one from the kinship table with existing records. If those are sorted on Id/Id2, it would be pretty easy to walk those and only do the import/update/delete for records that have actually changed.

[bbimber]

hey @labkey-martyp: can you confirm you're still good with this? you may not have seen, but I added a couple minor changes since the first approval. github will let me merge but doesnt seem in the spirit of review. note: these test failures to not seem related and the tests to exercise kinship seem to have passed.

[labkey-martyp]

@bbimber Your changes look good. Thanks. Yeah looks like those are just flaky test failures not related to this PR.