fix: schemas/public_.sql: sync_*(): use util.copy() instead of CREATE TABLE AS so that table and column comments are also copied. this avoids the need to separately add the same comments to the view and its materialized table.

bugfix: schemas/util.sql: recreate(): need to handle case where util.mk_drop_from_create() is NULL

bugfix: schemas/util.sql: mk_drop_from_create(): only match CREATE if no custom DROP came before it

bugfix: schemas/public_.sql: sync_geoscrub_input_to_view(): `CREATE TABLE geoscrub_input AS __`: needs `LIMIT 0`

fix: schemas/util.sql: explain2notice_msg_if_can(): also need to catch invalid_cursor_definition ("cannot open multi-query plan as cursor")

schemas/public_.sql: sync_analytical_stem_to_view(): removed DROP TABLE IF EXISTS because this is now done automatically by util.recreate()

schemas/util.sql: added copy()

schemas/util.sql: added copy_data()

fix: inputs/.TNRS/schema.sql: added back index on Name_submitted, which is needed for tnrs_input_name to work properly (now that there is no automatic index created by a unique constraint)

schemas/vegbien.sql: tnrs_input_name: don't scrub accepted names, as using multiple matches per name no longer provides a single accepted name to scrub. instead, the Accepted_* fields can be whitespace-split to generate the same columns that would have been generated by the scrubbing (and without the overhead of the extra TNRS call).