schemas/Makefile: removed no longer used rename/%, rotate. use %/install, %/publish instead to manage public schema versions.

schemas/Makefile: removed no longer needed public/uninstall, which is handled more generally by `%/uninstall: vegbien.sql` instead

schemas/Makefile: `%/uninstall: vegbien.sql`: added support for *_validation schema

schemas/Makefile: %/uninstall: split into 2 rules like for %/install, so that different actions can easily be added for the public schema

schemas/Makefile: %/publish: support auxiliary *_validation schemas

bugfix: schemas/Makefile: `%/install: vegbien.sql`: need -e before each s/// expression

schemas/Makefile: `%/install: vegbien.sql`: create a custom _validation schema for *each public schema, rather than one for the most recently-created public schema. this allows validations to continue to be run against a previous version of the DB while a new version is being imported.

schemas/vegbien.ERD.mwb: regenerated exports

schemas/vegbien.ERD.mwb: synced to schemas/vegbien.my.sql, which adds the validation schema

schemas/util.sql: added diff_views()

schemas/util.sql: added diff_cols(), for use with the aggregating validations

bugfix: schemas/Makefile: public/uninstall: need to uninstall the validation schema in addition to the public schema (deleting the public schema would only cascade to the contents of the validation schema)

schemas/vegbien.sql: added validation schema, which contains queries used in the aggregating validations. schemas/Makefile: include schemas that depend on `public` (eg. `validation`) so they are restored along with it.

fix: schemas/Makefile: %/uninstall: don't prompt to delete auxiliary schemas, as these do not contain data

schemas/vegbien.sql: added top_plot view

schemas/vegbien.sql: location_set_top_plot(): only do the lookup of the parent's top_plot when there is a parent_id

schemas/vegbien.sql: location: added autopopulated top_plot

schemas/vegbien.sql: locationevent: added locationevent_pull_forward_from_parent() trigger which populates project_id from the parent locationevent

schemas/vegbien.sql: locationevent__pull_forward_from_parent(): renamed to locationevent_pull_forward_from_parent_for_stratum() since it actually has a filter that causes it to apply only to stratum events

bugfix: schemas/vegbien.sql: commconcept: added missing unique constraint on commdescription, which is also an identifying field in addition to the ID (which is in commname.commname)

bugfix: schemas/vegbien.sql: locationevent__communities(): inner-join to commdetermination instead of left-joining, so that this does not produce community entries (which occur because there is one commclass per locationevent, but only some commclasses will then have a commdetermination)

schemas/vegbien.sql: projectcontributor: added missing unique constraint

moved everything into /trunk/ to create the standard svn layout, for use with tools that require this (eg. git-svn). IMPORTANT: do NOT do an `svn up`. instead, re-use your working copy's existing files with `svn switch` (http://svnbook.red-bean.com/en/1.6/svn.ref.svn.c.switch.html).

schemas/vegbien.sql: plot.**, analytical_stem_view, analytical_plot: added project_contributors

schemas/vegbien.sql: added project_contributors()

schemas/vegbien.sql: analytical_stem_view: scientificName_verbatim: don't use taxonverbatim.taxonname+author as the scientificName_verbatim if only the author is provided. (this lead to weird scientificName_verbatims that contain just the author.)

bugfix: schemas/vegbien.sql: location: use the place_id from the parent location when no place_id is specified. this fixes a bug in analytical_stem_view where the parent location's place_id was used because it was sometimes missing from the sublocation, but the parent place_id itself was sometimes missing instead if sublocations each had their own place information. this way, it is always available directly in the sublocation, populated from the parent location if needed.

bugfix: schemas/vegbien.sql: location: added place_id which is autopopulated from the current locationplace. join on this in plot.**, to avoid a 1:many join when a location has multiple locationplaces.

bugfix: schemas/vegbien.sql: locationevent_unique_within_parent_by_location unique index: need COALESCE around location_id since it's nullable

fix: schemas/vegbien.sql: analytical_specimen: added specimens-related columns that are in analytical_plot

bugfix: schemas/vegbien.sql: updated for renamed county_centroids column names

schemas/vegbien.sql: include the family_higher_plant_group lookup table values so that these don't need to be regenerated from the NCBI nodes whenever the DB is reloaded

schemas/vegbien.sql: taxonlabel_update_ancestors(): don't do an index scan if the value being scanned for is NULL, to support testing this function without the indexes in place, without extra full-table scans for NULL values affecting things. this can be used to determine if the function is actually using the indexes, by turning them off and seeing if the runtime changes.

schemas/util.sql: explain2table(): documented usage:
PERFORM util.explain2table($$
query
$$);

schemas/util.sql: explain2table(): by default, use the util.explain table

schemas/util.sql: added explain table

schemas/util.sql: added explain2notice()

schemas/util.sql: added explain2str()

schemas/util.sql: added explain2table()

schemas/util.sql: added explain()

schemas/vegbien.sql: taxonlabel_update_ancestors(): don't create a performance-intensive nested transaction (EXCEPTION block) for each INSERT, because there should no longer be duplicate ancestors, so it's OK to abort the whole transaction if this assertion fails

bugfix: schemas/vegbien.sql: taxonlabel_update_ancestors_on_{insert,update}(): only use either the matched taxon's ancestors or the parent's ancestors, to avoid issues related to duplication between these two ancestors lists. this also fixes a bug where the 2nd taxonlabel_update_ancestors() call assumes that the existing ancestors are for the old parent, when in fact they have actually just been set to those for the new matched taxon (which horribly confuses taxonlabel_update_ancestors()).

schemas/vegbien.sql: _taxonlabel_set_parent_id(): just use a plain UPDATE statement, to avoid the significant parsing and stringification overhead of EXECUTE and quote_nullable(). it is not clear that EXECUTE is actually necessary to avoid caching the query plan, because the cache should be invalidated automatically when the table's ANALYZE statistics are regenerated.

schemas/vegbien.sql: removed unused function _taxonlabel_set_matched_label_id(), which refers to obsolete fields

schemas/vegbien.sql: synced to DB (the view renderer apparently changed the text of a view)

bugfix: schemas/Makefile: %/uninstall: always confirm before removing an existing schema, not just for public and r*, because an auxiliary schema might also be used as $version and reinstalled by bin/import_all

schemas/vegbien.sql: analytical_stem_view: scrubbed_author: removed empty COALESCE around value (left over from when multiple values needed to be combined for many TNRS fields)

schemas/vegbien.sql: provider_count_view: source totals: use the much faster query developed for Brad (wiki.vegpath.org/VegBIEN_FAQ#from-Brad-on-2013-12-4), which avoids the need to do a GROUP BY on all of analytical_stem. eventually, we will want to apply the same optimization to the first publisher subtotals.

schemas/vegbien.ERD.mwb: regenerated exports

schemas/postgresql.conf: work_mem: documented that this seemingly small # is multiplied by max_connections, i.e. 256 MB * 100 = 26 GB, which approaches total memory (32 GB)

bugfix: schemas/Makefile: $(confirmRmPublicSchema): only prompt to delete the schema if it actually exists. this avoids prompting to remove a non-existent schema at the beginning of bin/import_all, which requires user attention. since bin/import_all is often run with a delayed start (e.g. to wait for a staging table reinstall to complete), the user may not be at the terminal when this message is displayed, and without this fix, the import would be prevented from running until they return.

schemas/vegbien.sql: collected_dates: documented runtime (2.5 min)

schemas/vegbien.sql: collected_date_min: replaced with collected_dates view that lists all dates we have, so that we can determine which of these may be valid. it turns out that we have data collected from very far back (to the year 1), which are not merely 2-digit years because PostgreSQL will only parse early years when there are 4 digits.

bugfix: schemas/vegbien.sql: collected_date_min: exclude invalid dates < 1000-01-01

bugfix: schemas/vegbien.sql: collected_date_min: exclude -infinity

schemas/vegbien.sql: added collected_date_min view

schemas/vegbien.sql: analytical_stem_view: added taxonomic_status. notice that PostgreSQL 9.3 puts each view column on a separate line, making it much easier to review the svn diff!

fix: schemas/vegbien.sql: taxon_trait_view: include only TNRS-valid names

schemas/vegbien.sql: sync_analytical_stem_to_view(): use new util.force_recreate() instead of manually dropping and re-creating every view that uses this. this avoids the need to add several lines to this function every time we add a new scientific view (of which we expect to have many), because force_recreate()'s error parsing handles this automatically. this makes it possible for a non-expert user to add scientific views without compromising the ability to add columns to analytical_stem_view, because they don't need to understand Postgres's dependency error messages when updating analytical_stem with this function.

schemas/util.sql: added force_recreate(), for use by sync_analytical_stem_to_view(). this uses the new `GET STACKED DIAGNOSTICS` in PostgreSQL 9.3 to access the DETAIL section of the dependent_objects_still_exist error.

schemas/*.conf: upgraded to PostgreSQL 9.3, which is needed for proper exception parsing in the auto-re-create-views functionality

bugfix: schemas/vegbien.sql: plantobservation_aggregateoccurrence_count_1(): only default aggregateoccurrence.count to 1 for specimens data, because plots data may have any number of individuals in a taxon_presence record that has no explicit individual_count

schemas/*.sql: updated for PostgreSQL 9.3. this reorders some functions, adds empty comment headers for omitted SEQUENCE SET commands, and (best of all) finally splits view columns onto multiple lines, so that changes in the columns are actually legible (and produce their own svn diff!)

Makefile, schemas/.Mac.conf: upgraded to PostgreSQL 9.3, which is needed for proper exception parsing in the auto-re-create-views functionality. this also removes the Mac 10.8 Mountain Lion quirks, such as renaming the postgres user to _postgres (which messed everything up, but is now back to normal).

schemas/util.sql: added save_drop_views()

schemas/util.sql: added is_empty(anyarray)

schemas/util.sql: added regexp_matches_group()

schemas/util.sql: show_create_view(): also include GRANT statements, which are necessary to fully re-create the view

schemas/util.sql: added show_grants_for(table_ regclass), for use by show_create_view()

schemas/util.sql: added save_drop_view()

schemas/util.sql: added show_create_view()

bugfix: schemas/vegbien.sql: scrubbed_morphospecies_binomial: only append the morphospecies suffix if there is not a scrubbed specific epithet

bugfix: schemas/vegbien.sql: scrubbed_morphospecies_binomial: only populate this from the component ranks; do not put a full taxon name in here if it would otherwise be NULL

bugfix: schemas/vegbien.sql: tnrs_input_name: MatchedTaxon self-join: must use a NOT NULL column for a proper anti-join. this unfortunately requires the more verbose LEFT JOIN ON syntax (which allows using the pkey as the NOT NULL column) instead of NATURAL LEFT JOIN (which requires using another column, which are all nullable)

schemas/vegbien.sql: tnrs_input_name: use plain UNION, which automatically removes duplicates, rather than UNION ALL with a manual EXCEPT-removal of rows in the first SELECT

schemas/vegbien.sql: tnrs_input_name: updated to use taxon_scrub.scrubbed_unique_taxon_name.*, to avoid further dependencies on AcceptedTaxon

schemas/vegbien.sql: taxon_trait_view: updated to use new taxon_scrub

schemas/vegbien.sql: analytical_stem_view: updated to use new taxon_scrub. this avoids the need to manually COALESCE every accepted* and matched* field, and makes the formulas much clearer

bugfix: schemas/vegbien.sql: analytical_stem_view: scrubbed_taxon_name_no_author, scrubbed_author: need to COALESCE these to the matched* when no accepted* is available

schemas/vegbien.sql: analytical_stem_view, etc.: renamed scrubbed fields with the scrubbed_* prefix, to clearly distinguish these from the equivalent fields for other taxon names

bugfix: schemas/vegbien.sql: analytical_stem_view: family, genus: need to COALESCE these to the matched* when no accepted* is available

bugfix: schemas/vegbien.sql: 2013-7-10.Naia.range_limiting_factors: filter by coordinateUncertaintyInMeters filter: assume true for rows with no coordinateUncertaintyInMeters

schemas/vegbien.sql: 2013-7-10.Naia.range_limiting_factors: filter by coordinateUncertaintyInMeters <= 10 km

schemas/vegbien.sql: 2013-7-10.Naia.range_limiting_factors: don't sort the results by occurrence_id, because this is not a meaningful ordering and prevents incremental output from the query

schemas/vegbien.sql: 2013-7-10.Naia.range_limiting_factors: also filter out rows without species

schemas/Makefile: $(confirmRmPublicSchema0): use "any ... schema" instead of "the ... schema" because the schema in question may not exist

fix: schemas/vegbien.sql: analytical_plot, analytical_specimen: removed derived columns that are not part of the validation

fix: schemas/vegbien.sql: analytical_plot, analytical_specimen: removed internal ID columns that are not part of the validation

schemas/vegbien.sql: analytical_plot: removed derived columns that should not be validated by data providers

schemas/vegbien.sql: analytical_specimen: synced to analytical_stem

schemas/vegbien.sql: analytical_plot: documented that this contains all of the analytical_stem columns, minus specimenHolderInstitutions, collection, accessionNumber, occurrenceID

schemas/vegbien.sql: analytical_plot: synced to analytical_stem

schemas/vegbien.sql: analytical_stem_view: added individualCount

schemas/vegbien.sql: plot.**, analytical_stem_view: added slopeAspect, slopeGradient

schemas/VegCore/ERD/VegCore.ERD.mwb: traceable.id_by_source: support multiple ids_by_source per traceable, because the same entity may be present in multiple datasources (e.g. if one got data from the other), and we would like to remove that duplicate

schemas/vegbien.sql: taxondetermination_set_iscurrent(): include new iscurrent__verbatim, so that taxondeterminations the datasource marks as current are always considered first. this currently applies to VegBank and CVS.

schemas/vegbien.sql: taxondetermination.isoriginal: made it nullable like iscurrent__verbatim, because this is populated from the datasource. taxondetermination_set_iscurrent() now supports isoriginal=NULL, so this is not a problem.

schemas/vegbien.sql: taxondetermination.is_datasource_current: renamed to iscurrent__verbatim and made it nullable, so that this can be used to store the verbatim iscurrent status

schemas/vegbien.sql: taxondetermination_set_iscurrent(): removed setting of is_datasource_current (which is now the same as iscurrent), so that this can be used to store the verbatim iscurrent status

schemas/vegbien.sql: taxondetermination_set_iscurrent(): isoriginal: make sure it is always either true or false, so that if the NOT NULL constraint on this is ever removed you don't end up with the incorrect sort order false, true, NULL (it should be false=NULL, true)