sql.py: put_table(), mk_select(): Switched back to having put_table() acquire the EXCLUSIVE locks, but right at the beginning of the transaction, in order to avoid lock upgrades which cause deadlocks

sql.py: with_autocommit(): Only allow turning autocommit on, because the opposite is not meaningful and may conflict with the session-global isolation level

sql.py: DbConn: Set the transaction isolation level to READ COMMITTED using set_isolation_level() so that the isolation level affects all transactions in the session, not just the current one

sql.py: DbConn: Always set the transaction isolation level to READ COMMITTED so that when a table is locked for update, its contents are frozen at that point rather than earlier. This ensures that no concurrent duplicate keys were inserted between the time the table was snapshotted (at the beginning of the transaction for SERIALIZABLE) and the time it was locked for update.

sql.py: put_table(): Removed locking output tables to prevent concurrent duplicate keys because that is now done automatically by mk_select()

sql.py: mk_select(): Filtering on no match: Lock the joined table in EXCLUSIVE mode to prevent concurrent duplicate keys when used with INSERT SELECT

sql_gen.py: Added underlying_table() and use it in underlying_col()

main Makefile: schemas/rotate: Fixed bug where needed to run schemas/public/install, not full schemas/install, after renaming public schema

sql.py: put_table(): Lock output tables to prevent concurrent duplicate keys

sql.py: Added lock_table()

bin/map: connect_db(): Only use autocommit mode if verbosity > 3, to avoid accidentally activating it if you want debug output in normal import mode

bin/map: connect_db(): Only use autocommit mode if verbosity > 2, because it causes the intermediate tables to be created as permanent tables, which you don't want unless you're actually debugging (verbosity = 2 is normal for column-based import)

sql.py: put_table(): remove_all_rows(): Changed log message to "Ignoring all rows" because NULL is not necessarily the pkey value that will be returned for the rows

sql.py: put_table(): Don't add index on columns that will have values filtered out, because indexes have already been added on all columns in the iteration's input table by flatten()

sql.py: DbConn._db(): Setting serializable isolation level: Always set this (if self.serializable is set), even in autocommit mode, because autocommit mode is implemented by manual commits in the DbConn wrapper object rather than using the underlying connection's autocommit mode (which does not allow setting the isolation level)

sql.py: DbConn._db(): Setting search_path: Use `SET search_path` and `SHOW search_path` instead of combining the old and new search_paths in SQL itself using `SELECT set_config('search_path', ...)`

csv2db: ProgressInputStream: Use default progress message 'Read %d line(s)' because there is not necessarily one CSV row per line, due to embedded newlines

input.Makefile: Staging tables: import/install-%: Only output to the log file if log option is non-empty (which it is by default)

csv2db: Support reinstalling just the errors table using new errors_table_only option

sql.py: Added drop_table()

schemas/vegbien.sql: method: Changed indexes to use `COALESCE` to match what sql_gen now does

schemas/vegbien.sql: specimenreplicate: Added indexes using COALESCE to match what sql_gen does

schemas/vegbien.sql: locationevent: Added indexes using COALESCE to match what sql_gen does

schemas/vegbien.ERD.mwb: Synced with schema

schemas/vegbien.sql: party: Changed indexes to use `COALESCE` to match what sql_gen now does

Wrap sys.stderr.write() calls in strings.to_raw_str() to avoid UnicodeEncodeErrors when stderr is to a file and the default encoding is ASCII

strings.py: Added to_raw_str()

bin/map: When logging the row # being processed, add 1 because row # is interally 0-based, but 1-based to the user

bin/map: Log the row # being processed with level=1.1 so that the user can see a status report if desired

exc.py: str_(): Fixed bug where UnicodeEncodeError would be raised when msg contains non-ASCII chars, by wrapping e.args⁰ in strings.ustr()

exc.py: print_ex(): Wrap msg in strings.to_unicode() to try to avoid UnicodeEncodeError when msg contains non-ASCII chars

sql.py: create_table(): Don't set pkey.nullable to False because the caller should make sure the pkey has the appropriate type

csv2db: Use sql_gen.TypedCol.nullable instead of manually adding 'NOT NULL' to the type. Ensure that pkeys are properly NOT NULL.

csv2db: Adding indexes: Create plain indexes using ensure_not_null=False because the indexes will primarily be used by the user to search for specific values, rather than by the mapping script which uses the ensure_not_null

sql.py: DbConn.col_info(): Run query with log_level=4 because it gathers information about database structure, and should have the same log_level as other queries that do that

csv2db: Adding indexes: Fixed bug where col.to_Col() could not be used because sql.add_index() does not support name-only columns (plain strings are OK, though)

sql.py: create_table(): has_pkey: Use new TypedCol.constraints to store 'PRIMARY KEY'

sql_gen.py: TypedCol: Added constraints instance var

sql_gen.py: EnsureNotNull: Made coalesce() all uppercase to match how pg_dump spells it

schemas/vegbien.sql: namedplace: Fixed bug where parent_id needed to be included in UNIQUE CONSTRAINT (now UNIQUE INDEX), since there can be more than one e.g. city of the same name if they are in different states

schemas/vegbien.sql: plantname: Changed indexes to use `COALESCE` to match what sql_gen now does

schemas/py_functions.sql: _dateRangeStart, _dateRangeEnd: Changed indexes to use `COALESCE` to match what sql_gen now does

schemas/py_functions.sql: _namePart: Changed indexes to use `COALESCE` to match what sql_gen now does

schemas/functions.sql: _nullIf: Changed indexes to use `COALESCE` to match what sql_gen now does

schemas/functions.sql: _nullIf: Require a non-NULL null-equivalent value

schemas/functions.sql: _label: Changed indexes to use `COALESCE` to match what sql_gen now does

schemas/functions.sql: _label: Require a non-NULL label

sql_gen.py: null_sentinels: Removed types where a sentinel doesn't make sense (unknown types, boolean) because types with no sentinel are now handled gracefully by users of ensure_not_null()

sql.py: add_index(): ensure_not_null: Handle unknown types gracefully

sql_gen.py: MockDb: Added col_info()

sql_gen.py: Use as_*() functions where the auto-wrapping was previously done manually

sql_gen.py: CompareCond.to_str(): Use ensure_not_null()'s new type_ param to apply same function to both sides but not if the right side is already NOT NULL

sql_gen.py: null_sentinels: Added value for character varying type

sql_gen.py: ensure_not_null(): Support non-column inputs if type_ is set

sql_gen.py: null_sentinels: Added value for USER-DEFINED type

sql.py: mk_select(): Joins: Filtering on no match: Use '~=' sql_gen.CompareCond operator so that IS NULL is always used, regardless of the not-null column's nullability

sql_gen.py: null_sentinels: Added value for boolean type

sql_gen.py: ensure_not_null(): Added type_ param to override the underlying column's type

sql_gen.py: EnsureNotNull: Take a type param instead of a null param so that the EnsureNotNull object stores the underlying column's type

sql_gen.py: underlying_col(): Support non-Col inputs

sql_gen.py: EnsureNotNull: Removed default value for null param to remind user that default value depends on value's type and will not always be a string

sql.py: add_index(): Added ensure_not_null param to disable the ensure_not_null functionality to force a plain index

sql.py: flatten(): Add indexes on the created table so its columns can be used in an O(n) merge join

sql_gen.py: null_sentinels: Added value for integer type

sql_gen.py: CompareCond.to_str(): Always wrap the left-side column if it's nullable. Wrap the right-side value if the left side was wrapped, rather than if both the left and right side are nullable. This causes coalesce() indexes to be used to look up NULL values using the value NULL gets coalesced to, rather than doing a sequential scan.

sql_gen.py: Run truncate() on all identifiers so that literal-string-based lookups for an identifier (such as in db.col_info()) don't use the untruncated value

sql_gen.py: Added truncate()

sql.py: put_table(): Resolving default value column: Fixed bug where default value of None was used as a key for mapping, even though this is an invalid Col name

sql_gen.py: ensure_not_null(): If input column cannot be ensured to be NULL, pass any raised exception through rather than suppressing it and leaving the column in a nullable state

schemas/functions.sql: _merge: Changed indexes to use `COALESCE` to match what sql_gen now does

schemas/functions.sql: _alt: Changed indexes to use `COALESCE` to match what sql_gen now does

sql_gen.py: CompareCond.to_str(): Handle nullable columns using ensure_not_null()

sql_gen.py: ensure_not_null(): Raise NoUnderlyingTableException if can't ensure not null for that reason

sql_gen.py: is_underlying_table(): Support non-Table inputs

sql_gen.py: NamedValues: Call set_cols_table() with the created table, not just the name, so that is_underlying_table() works properly

sql_gen.py: underlying_col(): If no underlying table, raise NoUnderlyingTableException

sql_gen.py: Added is_underlying_table()

sql_gen.py: ensure_not_null(): Call underlying_col() on the column to remove all renamings

sql_gen.py: Added underlying_col()

sql_gen.py: Join.to_str(): join(): Removed no longer needed `*_table = *_table.to_Table()`

sql_gen.py: Col: Support Table objects that are not just names, by calling `.to_Table()` on the table before stringifying it

sql_gen.py: Col: Support Table objects that are not just names, by calling `.to_Table()` on the table before stringifying it

sql_gen.py: ensure_not_null(): Added ignore_unknown_type param

sql_gen.py: CompareCond.to_str(): Put handling nullable columns as a separate step so it can be expanded

csv2db: Errors table: Removed no longer needed sql_gen.EnsureNotNull() because this is now added automatically

sql.py: add_index(): Handle nullable columns by using sql_gen.ensure_not_null() to automatically add a coalesce() wrapper where needed

sql_gen.py: Added ensure_not_null()

sql.py: DbConn.col_info(): Fixed bug where is_nullable needed to be cast to a boolean

sql.py: cast(): Support string column name inputs

sql.py: DbConn: Renamed col_default() to col_info() and have it return a sql_gen.TypedCol object containing all the TypedCol info about the column, not just the default value

sql_gen.py: TypedCol: Added default and nullable params

dicts.py: Import util after items that util depends on have been defined, to avoid unsatisfied circular dependency

sql.py: DbConn.col_default(): Pass the connection to sql_gen.as_Code() so it fixes the syntax on values returned by PostgreSQL

sql_gen.py: as_Code(): Added optional db param, which causes the function to run db.std_code() on the value to fix the syntax

sql.py: DbConn: Added std_code()

db_xml.py: Removed into_table_name() because this functionality is now handled by sql.into_table_name()

sql.py: into_table_name(): Also parse hierarchical tables (mappings with a rank column) using a special syntax

sql.py: put_table(): Fixed bug where distinct_on included columns that were not in the input table, and were thus incorrectly taken from the LEFT JOINed output table

sql.py: track_data_error(): Do nothing if cols are empty, because mk_track_data_error() requires at least one col. mk_track_data_error(): Assert that cols are not empty because VALUES clause requires at least one row.

bin/map: by_col: Pass on_error to db_xml.put_table() that calls ex_tracker.track()