Regex / Parse PostgreSQL names

Question

Going through PostgreSQL Names and Identifiers, I'm trying to figure out the following:

Is it possible to implement a regular expression (or parse it otherwise) to check whether a PostgreSQL name / identifier needs to be wrapped in double-quotes or not? And if possible, also to check if it is a Unicode-style name or not.

I need to inject an identifier name into an SQL query, and I would like to wrap this into double quotes only if the name needs it, because I read that adding quotes where they are not needed in fact changes the name format (see updates below).

UPDATE-1:

Ultimately, I need to implement the following function:

function prepareSQLName(name) {

    if (/* this is a regular name*/) {
        return name;
    }

    if (/* this is a unicode name */) {
        return 'U&"' + name + '"';
    } else {
        /* needs double quotes only */
        return '"' + name + '"';
    }
}

One of the major points of concern - processing non-English names. For example, the following is a perfectly valid PostgreSQL (in Russian) that doesn't require wrapping anything into double quotes:

create table жирафы(вес int, рост int, возраст int, имя text);
select имя, вес from жирафы order by рост;

UPDATE-2:

Examples of why it is such a big deal...

CREATE TABLE Test1(name text);

-- can be queried as:
SELECT * FROM Test1
SELECT * FROM test1
SELECT * FROM "test1"

-- but not as:
SELECT * FROM "Test1"

another one:

CREATE TABLE "Test2"(name text);

-- can be queried only as:
SELECT * FROM "Test2"

And it gets even trickier for Unicode-formatted names, ones that require U& in front.

Answer 1

var XRegExp = require('xregexp');

var isUnquotedIdentifier = new XRegExp('^[\pL_][\pL\pM_0-9$]*$');
var reservedWords = new Set(['A', 'ABORT', 'ABS', 'ABSENT', 'ABSOLUTE', 'ACCESS', 'ACCORDING', 'ACTION', 'ADA', 'ADD', 'ADMIN', 'AFTER', 'AGGREGATE', 'ALL', 'ALLOCATE', 'ALSO', 'ALTER', 'ALWAYS', 'ANALYSE', 'ANALYZE', 'AND', 'ANY', 'ARE', 'ARRAY', 'ARRAY_AGG', 'ARRAY_MAX_CARDINALITY', 'AS', 'ASC', 'ASENSITIVE', 'ASSERTION', 'ASSIGNMENT', 'ASYMMETRIC', 'AT', 'ATOMIC', 'ATTRIBUTE', 'ATTRIBUTES', 'AUTHORIZATION', 'AVG', 'BACKWARD', 'BASE64', 'BEFORE', 'BEGIN', 'BEGIN_FRAME', 'BEGIN_PARTITION', 'BERNOULLI', 'BETWEEN', 'BIGINT', 'BINARY', 'BIT', 'BIT_LENGTH', 'BLOB', 'BLOCKED', 'BOM', 'BOOLEAN', 'BOTH', 'BREADTH', 'BY', 'C', 'CACHE', 'CALL', 'CALLED', 'CARDINALITY', 'CASCADE', 'CASCADED', 'CASE', 'CAST', 'CATALOG', 'CATALOG_NAME', 'CEIL', 'CEILING', 'CHAIN', 'CHAR', 'CHARACTER', 'CHARACTERISTICS', 'CHARACTERS', 'CHARACTER_LENGTH', 'CHARACTER_SET_CATALOG', 'CHARACTER_SET_NAME', 'CHARACTER_SET_SCHEMA', 'CHAR_LENGTH', 'CHECK', 'CHECKPOINT', 'CLASS', 'CLASS_ORIGIN', 'CLOB', 'CLOSE', 'CLUSTER', 'COALESCE', 'COBOL', 'COLLATE', 'COLLATION', 'COLLATION_CATALOG', 'COLLATION_NAME', 'COLLATION_SCHEMA', 'COLLECT', 'COLUMN', 'COLUMNS', 'COLUMN_NAME', 'COMMAND_FUNCTION', 'COMMAND_FUNCTION_CODE', 'COMMENT', 'COMMENTS', 'COMMIT', 'COMMITTED', 'CONCURRENTLY', 'CONDITION', 'CONDITION_NUMBER', 'CONFIGURATION', 'CONFLICT', 'CONNECT', 'CONNECTION', 'CONNECTION_NAME', 'CONSTRAINT', 'CONSTRAINTS', 'CONSTRAINT_CATALOG', 'CONSTRAINT_NAME', 'CONSTRAINT_SCHEMA', 'CONSTRUCTOR', 'CONTAINS', 'CONTENT', 'CONTINUE', 'CONTROL', 'CONVERSION', 'CONVERT', 'COPY', 'CORR', 'CORRESPONDING', 'COST', 'COUNT', 'COVAR_POP', 'COVAR_SAMP', 'CREATE', 'CROSS', 'CSV', 'CUBE', 'CUME_DIST', 'CURRENT', 'CURRENT_CATALOG', 'CURRENT_DATE', 'CURRENT_DEFAULT_TRANSFORM_GROUP', 'CURRENT_PATH', 'CURRENT_ROLE', 'CURRENT_ROW', 'CURRENT_SCHEMA', 'CURRENT_TIME', 'CURRENT_TIMESTAMP', 'CURRENT_TRANSFORM_GROUP_FOR_TYPE', 'CURRENT_USER', 'CURSOR', 'CURSOR_NAME', 'CYCLE', 'DATA', 'DATABASE', 'DATALINK', 'DATE', 'DATETIME_INTERVAL_CODE', 'DATETIME_INTERVAL_PRECISION', 'DAY', 'DB', 'DEALLOCATE', 'DEC', 'DECIMAL', 'DECLARE', 'DEFAULT', 'DEFAULTS', 'DEFERRABLE', 'DEFERRED', 'DEFINED', 'DEFINER', 'DEGREE', 'DELETE', 'DELIMITER', 'DELIMITERS', 'DENSE_RANK', 'DEPTH', 'DEREF', 'DERIVED', 'DESC', 'DESCRIBE', 'DESCRIPTOR', 'DETERMINISTIC', 'DIAGNOSTICS', 'DICTIONARY', 'DISABLE', 'DISCARD', 'DISCONNECT', 'DISPATCH', 'DISTINCT', 'DLNEWCOPY', 'DLPREVIOUSCOPY', 'DLURLCOMPLETE', 'DLURLCOMPLETEONLY', 'DLURLCOMPLETEWRITE', 'DLURLPATH', 'DLURLPATHONLY', 'DLURLPATHWRITE', 'DLURLSCHEME', 'DLURLSERVER', 'DLVALUE', 'DO', 'DOCUMENT', 'DOMAIN', 'DOUBLE', 'DROP', 'DYNAMIC', 'DYNAMIC_FUNCTION', 'DYNAMIC_FUNCTION_CODE', 'EACH', 'ELEMENT', 'ELSE', 'EMPTY', 'ENABLE', 'ENCODING', 'ENCRYPTED', 'END', 'END-EXEC', 'END_FRAME', 'END_PARTITION', 'ENFORCED', 'ENUM', 'EQUALS', 'ESCAPE', 'EVENT', 'EVERY', 'EXCEPT', 'EXCEPTION', 'EXCLUDE', 'EXCLUDING', 'EXCLUSIVE', 'EXEC', 'EXECUTE', 'EXISTS', 'EXP', 'EXPLAIN', 'EXPRESSION', 'EXTENSION', 'EXTERNAL', 'EXTRACT', 'FALSE', 'FAMILY', 'FETCH', 'FILE', 'FILTER', 'FINAL', 'FIRST', 'FIRST_VALUE', 'FLAG', 'FLOAT', 'FLOOR', 'FOLLOWING', 'FOR', 'FORCE', 'FOREIGN', 'FORTRAN', 'FORWARD', 'FOUND', 'FRAME_ROW', 'FREE', 'FREEZE', 'FROM', 'FS', 'FULL', 'FUNCTION', 'FUNCTIONS', 'FUSION', 'G', 'GENERAL', 'GENERATED', 'GET', 'GLOBAL', 'GO', 'GOTO', 'GRANT', 'GRANTED', 'GREATEST', 'GROUP', 'GROUPING', 'GROUPS', 'HANDLER', 'HAVING', 'HEADER', 'HEX', 'HIERARCHY', 'HOLD', 'HOUR', 'ID', 'IDENTITY', 'IF', 'IGNORE', 'ILIKE', 'IMMEDIATE', 'IMMEDIATELY', 'IMMUTABLE', 'IMPLEMENTATION', 'IMPLICIT', 'IMPORT', 'IN', 'INCLUDING', 'INCREMENT', 'INDENT', 'INDEX', 'INDEXES', 'INDICATOR', 'INHERIT', 'INHERITS', 'INITIALLY', 'INLINE', 'INNER', 'INOUT', 'INPUT', 'INSENSITIVE', 'INSERT', 'INSTANCE', 'INSTANTIABLE', 'INSTEAD', 'INT', 'INTEGER', 'INTEGRITY', 'INTERSECT', 'INTERSECTION', 'INTERVAL', 'INTO', 'INVOKER', 'IS', 'ISNULL', 'ISOLATION', 'JOIN', 'K', 'KEY', 'KEY_MEMBER', 'KEY_TYPE', 'LABEL', 'LAG', 'LANGUAGE', 'LARGE', 'LAST', 'LAST_VALUE', 'LATERAL', 'LEAD', 'LEADING', 'LEAKPROOF', 'LEAST', 'LEFT', 'LENGTH', 'LEVEL', 'LIBRARY', 'LIKE', 'LIKE_REGEX', 'LIMIT', 'LINK', 'LISTEN', 'LN', 'LOAD', 'LOCAL', 'LOCALTIME', 'LOCALTIMESTAMP', 'LOCATION', 'LOCATOR', 'LOCK', 'LOCKED', 'LOGGED', 'LOWER', 'M', 'MAP', 'MAPPING', 'MATCH', 'MATCHED', 'MATERIALIZED', 'MAX', 'MAXVALUE', 'MAX_CARDINALITY', 'MEMBER', 'MERGE', 'MESSAGE_LENGTH', 'MESSAGE_OCTET_LENGTH', 'MESSAGE_TEXT', 'METHOD', 'MIN', 'MINUTE', 'MINVALUE', 'MOD', 'MODE', 'MODIFIES', 'MODULE', 'MONTH', 'MORE', 'MOVE', 'MULTISET', 'MUMPS', 'NAME', 'NAMES', 'NAMESPACE', 'NATIONAL', 'NATURAL', 'NCHAR', 'NCLOB', 'NESTING', 'NEW', 'NEXT', 'NFC', 'NFD', 'NFKC', 'NFKD', 'NIL', 'NO', 'NONE', 'NORMALIZE', 'NORMALIZED', 'NOT', 'NOTHING', 'NOTIFY', 'NOTNULL', 'NOWAIT', 'NTH_VALUE', 'NTILE', 'NULL', 'NULLABLE', 'NULLIF', 'NULLS', 'NUMBER', 'NUMERIC', 'OBJECT', 'OCCURRENCES_REGEX', 'OCTETS', 'OCTET_LENGTH', 'OF', 'OFF', 'OFFSET', 'OIDS', 'OLD', 'ON', 'ONLY', 'OPEN', 'OPERATOR', 'OPTION', 'OPTIONS', 'OR', 'ORDER', 'ORDERING', 'ORDINALITY', 'OTHERS', 'OUT', 'OUTER', 'OUTPUT', 'OVER', 'OVERLAPS', 'OVERLAY', 'OVERRIDING', 'OWNED', 'OWNER', 'P', 'PAD', 'PARAMETER', 'PARAMETER_MODE', 'PARAMETER_NAME', 'PARAMETER_ORDINAL_POSITION', 'PARAMETER_SPECIFIC_CATALOG', 'PARAMETER_SPECIFIC_NAME', 'PARAMETER_SPECIFIC_SCHEMA', 'PARSER', 'PARTIAL', 'PARTITION', 'PASCAL', 'PASSING', 'PASSTHROUGH', 'PASSWORD', 'PATH', 'PERCENT', 'PERCENTILE_CONT', 'PERCENTILE_DISC', 'PERCENT_RANK', 'PERIOD', 'PERMISSION', 'PLACING', 'PLANS', 'PLI', 'POLICY', 'PORTION', 'POSITION', 'POSITION_REGEX', 'POWER', 'PRECEDES', 'PRECEDING', 'PRECISION', 'PREPARE', 'PREPARED', 'PRESERVE', 'PRIMARY', 'PRIOR', 'PRIVILEGES', 'PROCEDURAL', 'PROCEDURE', 'PROGRAM', 'PUBLIC', 'QUOTE', 'RANGE', 'RANK', 'READ', 'READS', 'REAL', 'REASSIGN', 'RECHECK', 'RECOVERY', 'RECURSIVE', 'REF', 'REFERENCES', 'REFERENCING', 'REFRESH', 'REGR_AVGX', 'REGR_AVGY', 'REGR_COUNT', 'REGR_INTERCEPT', 'REGR_R2', 'REGR_SLOPE', 'REGR_SXX', 'REGR_SXY', 'REGR_SYY', 'REINDEX', 'RELATIVE', 'RELEASE', 'RENAME', 'REPEATABLE', 'REPLACE', 'REPLICA', 'REQUIRING', 'RESET', 'RESPECT', 'RESTART', 'RESTORE', 'RESTRICT', 'RESULT', 'RETURN', 'RETURNED_CARDINALITY', 'RETURNED_LENGTH', 'RETURNED_OCTET_LENGTH', 'RETURNED_SQLSTATE', 'RETURNING', 'RETURNS', 'REVOKE', 'RIGHT', 'ROLE', 'ROLLBACK', 'ROLLUP', 'ROUTINE', 'ROUTINE_CATALOG', 'ROUTINE_NAME', 'ROUTINE_SCHEMA', 'ROW', 'ROWS', 'ROW_COUNT', 'ROW_NUMBER', 'RULE', 'SAVEPOINT', 'SCALE', 'SCHEMA', 'SCHEMA_NAME', 'SCOPE', 'SCOPE_CATALOG', 'SCOPE_NAME', 'SCOPE_SCHEMA', 'SCROLL', 'SEARCH', 'SECOND', 'SECTION', 'SECURITY', 'SELECT', 'SELECTIVE', 'SELF', 'SENSITIVE', 'SEQUENCE', 'SEQUENCES', 'SERIALIZABLE', 'SERVER', 'SERVER_NAME', 'SESSION', 'SESSION_USER', 'SET', 'SETOF', 'SETS', 'SHARE', 'SHOW', 'SIMILAR', 'SIMPLE', 'SIZE', 'SKIP', 'SMALLINT', 'SNAPSHOT', 'SOME', 'SOURCE', 'SPACE', 'SPECIFIC', 'SPECIFICTYPE', 'SPECIFIC_NAME', 'SQL', 'SQLCODE', 'SQLERROR', 'SQLEXCEPTION', 'SQLSTATE', 'SQLWARNING', 'SQRT', 'STABLE', 'STANDALONE', 'START', 'STATE', 'STATEMENT', 'STATIC', 'STATISTICS', 'STDDEV_POP', 'STDDEV_SAMP', 'STDIN', 'STDOUT', 'STORAGE', 'STRICT', 'STRIP', 'STRUCTURE', 'STYLE', 'SUBCLASS_ORIGIN', 'SUBMULTISET', 'SUBSTRING', 'SUBSTRING_REGEX', 'SUCCEEDS', 'SUM', 'SYMMETRIC', 'SYSID', 'SYSTEM', 'SYSTEM_TIME', 'SYSTEM_USER', 'T', 'TABLE', 'TABLES', 'TABLESAMPLE', 'TABLESPACE', 'TABLE_NAME', 'TEMP', 'TEMPLATE', 'TEMPORARY', 'TEXT', 'THEN', 'TIES', 'TIME', 'TIMESTAMP', 'TIMEZONE_HOUR', 'TIMEZONE_MINUTE', 'TO', 'TOKEN', 'TOP_LEVEL_COUNT', 'TRAILING', 'TRANSACTION', 'TRANSACTIONS_COMMITTED', 'TRANSACTIONS_ROLLED_BACK', 'TRANSACTION_ACTIVE', 'TRANSFORM', 'TRANSFORMS', 'TRANSLATE', 'TRANSLATE_REGEX', 'TRANSLATION', 'TREAT', 'TRIGGER', 'TRIGGER_CATALOG', 'TRIGGER_NAME', 'TRIGGER_SCHEMA', 'TRIM', 'TRIM_ARRAY', 'TRUE', 'TRUNCATE', 'TRUSTED', 'TYPE', 'TYPES', 'UESCAPE', 'UNBOUNDED', 'UNCOMMITTED', 'UNDER', 'UNENCRYPTED', 'UNION', 'UNIQUE', 'UNKNOWN', 'UNLINK', 'UNLISTEN', 'UNLOGGED', 'UNNAMED', 'UNNEST', 'UNTIL', 'UNTYPED', 'UPDATE', 'UPPER', 'URI', 'USAGE', 'USER', 'USER_DEFINED_TYPE_CATALOG', 'USER_DEFINED_TYPE_CODE', 'USER_DEFINED_TYPE_NAME', 'USER_DEFINED_TYPE_SCHEMA', 'USING', 'VACUUM', 'VALID', 'VALIDATE', 'VALIDATOR', 'VALUE', 'VALUES', 'VALUE_OF', 'VARBINARY', 'VARCHAR', 'VARIADIC', 'VARYING', 'VAR_POP', 'VAR_SAMP', 'VERBOSE', 'VERSION', 'VERSIONING', 'VIEW', 'VIEWS', 'VOLATILE', 'WHEN', 'WHENEVER', 'WHERE', 'WHITESPACE', 'WIDTH_BUCKET', 'WINDOW', 'WITH', 'WITHIN', 'WITHOUT', 'WORK', 'WRAPPER', 'WRITE', 'XML', 'XMLAGG', 'XMLATTRIBUTES', 'XMLBINARY', 'XMLCAST', 'XMLCOMMENT', 'XMLCONCAT', 'XMLDECLARATION', 'XMLDOCUMENT', 'XMLELEMENT', 'XMLEXISTS', 'XMLFOREST', 'XMLITERATE', 'XMLNAMESPACES', 'XMLPARSE', 'XMLPI', 'XMLQUERY', 'XMLROOT', 'XMLSCHEMA', 'XMLSERIALIZE', 'XMLTABLE', 'XMLTEXT', 'XMLVALIDATE', 'YEAR', 'YES', 'ZONE']);
function prepareSQLName(name) {
    if (isUnquotedIdentifier.test(name) && !reservedWords.has(name.toUpperCase()))
        return name;

    // Let's quote name:
    return '"' + name.replace(/"/g, '""') + '"';
}

console.log(prepareSQLName('simple')); //simple
console.log(prepareSQLName('to "quote')); //"to ""quote"
console.log(prepareSQLName('простий')); //простий
console.log(prepareSQLName('в "лапки')); //"в ""лапки"

Notes:

1. We need xregexp module, because built-in RegExp class doesn't support testing of Unicode categories (\pL a.k.a. p{L} and \pM a.k.a \p{M}). Still, detecting letters is not 100%-clear (see "Details on detecting letters").
2. List of key words is taken from here. Using this list may be non-ideal solution, as not all of these identifiers are really prohibited for use in all possible cases (i.e. sometimes it may quote table name that is allowed to be unquoted, e.g. abs). On the other hand, this task cannot be solved ideally without adding namePurpose second parameter to prepareSQLName function (as some of these identifiers are allowed for use in specific cases only).

3. You do not need special handling of non-ASCII, if you want non-English letters to remain non-escaped (they may be non-escaped as in SELECT "ім'я" FROM користувачі). The only thing to be checked is that identifiers containing something other than letters/digits/_/$ should be quoted (as "ім'я"). But you may also additionally escape all non-ASCII characters, if you wish:

   var isUnquotedIdentifier = new RegExp('^[A-Za-z_][A-Za-z_0-9$]*$');
   var reservedWords = new Set(['A', 'ABORT', 'ABS', 'ABSENT', 'ABSOLUTE', 'ACCESS', 'ACCORDING', 'ACTION', 'ADA', 'ADD', 'ADMIN', 'AFTER', 'AGGREGATE', 'ALL', 'ALLOCATE', 'ALSO', 'ALTER', 'ALWAYS', 'ANALYSE', 'ANALYZE', 'AND', 'ANY', 'ARE', 'ARRAY', 'ARRAY_AGG', 'ARRAY_MAX_CARDINALITY', 'AS', 'ASC', 'ASENSITIVE', 'ASSERTION', 'ASSIGNMENT', 'ASYMMETRIC', 'AT', 'ATOMIC', 'ATTRIBUTE', 'ATTRIBUTES', 'AUTHORIZATION', 'AVG', 'BACKWARD', 'BASE64', 'BEFORE', 'BEGIN', 'BEGIN_FRAME', 'BEGIN_PARTITION', 'BERNOULLI', 'BETWEEN', 'BIGINT', 'BINARY', 'BIT', 'BIT_LENGTH', 'BLOB', 'BLOCKED', 'BOM', 'BOOLEAN', 'BOTH', 'BREADTH', 'BY', 'C', 'CACHE', 'CALL', 'CALLED', 'CARDINALITY', 'CASCADE', 'CASCADED', 'CASE', 'CAST', 'CATALOG', 'CATALOG_NAME', 'CEIL', 'CEILING', 'CHAIN', 'CHAR', 'CHARACTER', 'CHARACTERISTICS', 'CHARACTERS', 'CHARACTER_LENGTH', 'CHARACTER_SET_CATALOG', 'CHARACTER_SET_NAME', 'CHARACTER_SET_SCHEMA', 'CHAR_LENGTH', 'CHECK', 'CHECKPOINT', 'CLASS', 'CLASS_ORIGIN', 'CLOB', 'CLOSE', 'CLUSTER', 'COALESCE', 'COBOL', 'COLLATE', 'COLLATION', 'COLLATION_CATALOG', 'COLLATION_NAME', 'COLLATION_SCHEMA', 'COLLECT', 'COLUMN', 'COLUMNS', 'COLUMN_NAME', 'COMMAND_FUNCTION', 'COMMAND_FUNCTION_CODE', 'COMMENT', 'COMMENTS', 'COMMIT', 'COMMITTED', 'CONCURRENTLY', 'CONDITION', 'CONDITION_NUMBER', 'CONFIGURATION', 'CONFLICT', 'CONNECT', 'CONNECTION', 'CONNECTION_NAME', 'CONSTRAINT', 'CONSTRAINTS', 'CONSTRAINT_CATALOG', 'CONSTRAINT_NAME', 'CONSTRAINT_SCHEMA', 'CONSTRUCTOR', 'CONTAINS', 'CONTENT', 'CONTINUE', 'CONTROL', 'CONVERSION', 'CONVERT', 'COPY', 'CORR', 'CORRESPONDING', 'COST', 'COUNT', 'COVAR_POP', 'COVAR_SAMP', 'CREATE', 'CROSS', 'CSV', 'CUBE', 'CUME_DIST', 'CURRENT', 'CURRENT_CATALOG', 'CURRENT_DATE', 'CURRENT_DEFAULT_TRANSFORM_GROUP', 'CURRENT_PATH', 'CURRENT_ROLE', 'CURRENT_ROW', 'CURRENT_SCHEMA', 'CURRENT_TIME', 'CURRENT_TIMESTAMP', 'CURRENT_TRANSFORM_GROUP_FOR_TYPE', 'CURRENT_USER', 'CURSOR', 'CURSOR_NAME', 'CYCLE', 'DATA', 'DATABASE', 'DATALINK', 'DATE', 'DATETIME_INTERVAL_CODE', 'DATETIME_INTERVAL_PRECISION', 'DAY', 'DB', 'DEALLOCATE', 'DEC', 'DECIMAL', 'DECLARE', 'DEFAULT', 'DEFAULTS', 'DEFERRABLE', 'DEFERRED', 'DEFINED', 'DEFINER', 'DEGREE', 'DELETE', 'DELIMITER', 'DELIMITERS', 'DENSE_RANK', 'DEPTH', 'DEREF', 'DERIVED', 'DESC', 'DESCRIBE', 'DESCRIPTOR', 'DETERMINISTIC', 'DIAGNOSTICS', 'DICTIONARY', 'DISABLE', 'DISCARD', 'DISCONNECT', 'DISPATCH', 'DISTINCT', 'DLNEWCOPY', 'DLPREVIOUSCOPY', 'DLURLCOMPLETE', 'DLURLCOMPLETEONLY', 'DLURLCOMPLETEWRITE', 'DLURLPATH', 'DLURLPATHONLY', 'DLURLPATHWRITE', 'DLURLSCHEME', 'DLURLSERVER', 'DLVALUE', 'DO', 'DOCUMENT', 'DOMAIN', 'DOUBLE', 'DROP', 'DYNAMIC', 'DYNAMIC_FUNCTION', 'DYNAMIC_FUNCTION_CODE', 'EACH', 'ELEMENT', 'ELSE', 'EMPTY', 'ENABLE', 'ENCODING', 'ENCRYPTED', 'END', 'END-EXEC', 'END_FRAME', 'END_PARTITION', 'ENFORCED', 'ENUM', 'EQUALS', 'ESCAPE', 'EVENT', 'EVERY', 'EXCEPT', 'EXCEPTION', 'EXCLUDE', 'EXCLUDING', 'EXCLUSIVE', 'EXEC', 'EXECUTE', 'EXISTS', 'EXP', 'EXPLAIN', 'EXPRESSION', 'EXTENSION', 'EXTERNAL', 'EXTRACT', 'FALSE', 'FAMILY', 'FETCH', 'FILE', 'FILTER', 'FINAL', 'FIRST', 'FIRST_VALUE', 'FLAG', 'FLOAT', 'FLOOR', 'FOLLOWING', 'FOR', 'FORCE', 'FOREIGN', 'FORTRAN', 'FORWARD', 'FOUND', 'FRAME_ROW', 'FREE', 'FREEZE', 'FROM', 'FS', 'FULL', 'FUNCTION', 'FUNCTIONS', 'FUSION', 'G', 'GENERAL', 'GENERATED', 'GET', 'GLOBAL', 'GO', 'GOTO', 'GRANT', 'GRANTED', 'GREATEST', 'GROUP', 'GROUPING', 'GROUPS', 'HANDLER', 'HAVING', 'HEADER', 'HEX', 'HIERARCHY', 'HOLD', 'HOUR', 'ID', 'IDENTITY', 'IF', 'IGNORE', 'ILIKE', 'IMMEDIATE', 'IMMEDIATELY', 'IMMUTABLE', 'IMPLEMENTATION', 'IMPLICIT', 'IMPORT', 'IN', 'INCLUDING', 'INCREMENT', 'INDENT', 'INDEX', 'INDEXES', 'INDICATOR', 'INHERIT', 'INHERITS', 'INITIALLY', 'INLINE', 'INNER', 'INOUT', 'INPUT', 'INSENSITIVE', 'INSERT', 'INSTANCE', 'INSTANTIABLE', 'INSTEAD', 'INT', 'INTEGER', 'INTEGRITY', 'INTERSECT', 'INTERSECTION', 'INTERVAL', 'INTO', 'INVOKER', 'IS', 'ISNULL', 'ISOLATION', 'JOIN', 'K', 'KEY', 'KEY_MEMBER', 'KEY_TYPE', 'LABEL', 'LAG', 'LANGUAGE', 'LARGE', 'LAST', 'LAST_VALUE', 'LATERAL', 'LEAD', 'LEADING', 'LEAKPROOF', 'LEAST', 'LEFT', 'LENGTH', 'LEVEL', 'LIBRARY', 'LIKE', 'LIKE_REGEX', 'LIMIT', 'LINK', 'LISTEN', 'LN', 'LOAD', 'LOCAL', 'LOCALTIME', 'LOCALTIMESTAMP', 'LOCATION', 'LOCATOR', 'LOCK', 'LOCKED', 'LOGGED', 'LOWER', 'M', 'MAP', 'MAPPING', 'MATCH', 'MATCHED', 'MATERIALIZED', 'MAX', 'MAXVALUE', 'MAX_CARDINALITY', 'MEMBER', 'MERGE', 'MESSAGE_LENGTH', 'MESSAGE_OCTET_LENGTH', 'MESSAGE_TEXT', 'METHOD', 'MIN', 'MINUTE', 'MINVALUE', 'MOD', 'MODE', 'MODIFIES', 'MODULE', 'MONTH', 'MORE', 'MOVE', 'MULTISET', 'MUMPS', 'NAME', 'NAMES', 'NAMESPACE', 'NATIONAL', 'NATURAL', 'NCHAR', 'NCLOB', 'NESTING', 'NEW', 'NEXT', 'NFC', 'NFD', 'NFKC', 'NFKD', 'NIL', 'NO', 'NONE', 'NORMALIZE', 'NORMALIZED', 'NOT', 'NOTHING', 'NOTIFY', 'NOTNULL', 'NOWAIT', 'NTH_VALUE', 'NTILE', 'NULL', 'NULLABLE', 'NULLIF', 'NULLS', 'NUMBER', 'NUMERIC', 'OBJECT', 'OCCURRENCES_REGEX', 'OCTETS', 'OCTET_LENGTH', 'OF', 'OFF', 'OFFSET', 'OIDS', 'OLD', 'ON', 'ONLY', 'OPEN', 'OPERATOR', 'OPTION', 'OPTIONS', 'OR', 'ORDER', 'ORDERING', 'ORDINALITY', 'OTHERS', 'OUT', 'OUTER', 'OUTPUT', 'OVER', 'OVERLAPS', 'OVERLAY', 'OVERRIDING', 'OWNED', 'OWNER', 'P', 'PAD', 'PARAMETER', 'PARAMETER_MODE', 'PARAMETER_NAME', 'PARAMETER_ORDINAL_POSITION', 'PARAMETER_SPECIFIC_CATALOG', 'PARAMETER_SPECIFIC_NAME', 'PARAMETER_SPECIFIC_SCHEMA', 'PARSER', 'PARTIAL', 'PARTITION', 'PASCAL', 'PASSING', 'PASSTHROUGH', 'PASSWORD', 'PATH', 'PERCENT', 'PERCENTILE_CONT', 'PERCENTILE_DISC', 'PERCENT_RANK', 'PERIOD', 'PERMISSION', 'PLACING', 'PLANS', 'PLI', 'POLICY', 'PORTION', 'POSITION', 'POSITION_REGEX', 'POWER', 'PRECEDES', 'PRECEDING', 'PRECISION', 'PREPARE', 'PREPARED', 'PRESERVE', 'PRIMARY', 'PRIOR', 'PRIVILEGES', 'PROCEDURAL', 'PROCEDURE', 'PROGRAM', 'PUBLIC', 'QUOTE', 'RANGE', 'RANK', 'READ', 'READS', 'REAL', 'REASSIGN', 'RECHECK', 'RECOVERY', 'RECURSIVE', 'REF', 'REFERENCES', 'REFERENCING', 'REFRESH', 'REGR_AVGX', 'REGR_AVGY', 'REGR_COUNT', 'REGR_INTERCEPT', 'REGR_R2', 'REGR_SLOPE', 'REGR_SXX', 'REGR_SXY', 'REGR_SYY', 'REINDEX', 'RELATIVE', 'RELEASE', 'RENAME', 'REPEATABLE', 'REPLACE', 'REPLICA', 'REQUIRING', 'RESET', 'RESPECT', 'RESTART', 'RESTORE', 'RESTRICT', 'RESULT', 'RETURN', 'RETURNED_CARDINALITY', 'RETURNED_LENGTH', 'RETURNED_OCTET_LENGTH', 'RETURNED_SQLSTATE', 'RETURNING', 'RETURNS', 'REVOKE', 'RIGHT', 'ROLE', 'ROLLBACK', 'ROLLUP', 'ROUTINE', 'ROUTINE_CATALOG', 'ROUTINE_NAME', 'ROUTINE_SCHEMA', 'ROW', 'ROWS', 'ROW_COUNT', 'ROW_NUMBER', 'RULE', 'SAVEPOINT', 'SCALE', 'SCHEMA', 'SCHEMA_NAME', 'SCOPE', 'SCOPE_CATALOG', 'SCOPE_NAME', 'SCOPE_SCHEMA', 'SCROLL', 'SEARCH', 'SECOND', 'SECTION', 'SECURITY', 'SELECT', 'SELECTIVE', 'SELF', 'SENSITIVE', 'SEQUENCE', 'SEQUENCES', 'SERIALIZABLE', 'SERVER', 'SERVER_NAME', 'SESSION', 'SESSION_USER', 'SET', 'SETOF', 'SETS', 'SHARE', 'SHOW', 'SIMILAR', 'SIMPLE', 'SIZE', 'SKIP', 'SMALLINT', 'SNAPSHOT', 'SOME', 'SOURCE', 'SPACE', 'SPECIFIC', 'SPECIFICTYPE', 'SPECIFIC_NAME', 'SQL', 'SQLCODE', 'SQLERROR', 'SQLEXCEPTION', 'SQLSTATE', 'SQLWARNING', 'SQRT', 'STABLE', 'STANDALONE', 'START', 'STATE', 'STATEMENT', 'STATIC', 'STATISTICS', 'STDDEV_POP', 'STDDEV_SAMP', 'STDIN', 'STDOUT', 'STORAGE', 'STRICT', 'STRIP', 'STRUCTURE', 'STYLE', 'SUBCLASS_ORIGIN', 'SUBMULTISET', 'SUBSTRING', 'SUBSTRING_REGEX', 'SUCCEEDS', 'SUM', 'SYMMETRIC', 'SYSID', 'SYSTEM', 'SYSTEM_TIME', 'SYSTEM_USER', 'T', 'TABLE', 'TABLES', 'TABLESAMPLE', 'TABLESPACE', 'TABLE_NAME', 'TEMP', 'TEMPLATE', 'TEMPORARY', 'TEXT', 'THEN', 'TIES', 'TIME', 'TIMESTAMP', 'TIMEZONE_HOUR', 'TIMEZONE_MINUTE', 'TO', 'TOKEN', 'TOP_LEVEL_COUNT', 'TRAILING', 'TRANSACTION', 'TRANSACTIONS_COMMITTED', 'TRANSACTIONS_ROLLED_BACK', 'TRANSACTION_ACTIVE', 'TRANSFORM', 'TRANSFORMS', 'TRANSLATE', 'TRANSLATE_REGEX', 'TRANSLATION', 'TREAT', 'TRIGGER', 'TRIGGER_CATALOG', 'TRIGGER_NAME', 'TRIGGER_SCHEMA', 'TRIM', 'TRIM_ARRAY', 'TRUE', 'TRUNCATE', 'TRUSTED', 'TYPE', 'TYPES', 'UESCAPE', 'UNBOUNDED', 'UNCOMMITTED', 'UNDER', 'UNENCRYPTED', 'UNION', 'UNIQUE', 'UNKNOWN', 'UNLINK', 'UNLISTEN', 'UNLOGGED', 'UNNAMED', 'UNNEST', 'UNTIL', 'UNTYPED', 'UPDATE', 'UPPER', 'URI', 'USAGE', 'USER', 'USER_DEFINED_TYPE_CATALOG', 'USER_DEFINED_TYPE_CODE', 'USER_DEFINED_TYPE_NAME', 'USER_DEFINED_TYPE_SCHEMA', 'USING', 'VACUUM', 'VALID', 'VALIDATE', 'VALIDATOR', 'VALUE', 'VALUES', 'VALUE_OF', 'VARBINARY', 'VARCHAR', 'VARIADIC', 'VARYING', 'VAR_POP', 'VAR_SAMP', 'VERBOSE', 'VERSION', 'VERSIONING', 'VIEW', 'VIEWS', 'VOLATILE', 'WHEN', 'WHENEVER', 'WHERE', 'WHITESPACE', 'WIDTH_BUCKET', 'WINDOW', 'WITH', 'WITHIN', 'WITHOUT', 'WORK', 'WRAPPER', 'WRITE', 'XML', 'XMLAGG', 'XMLATTRIBUTES', 'XMLBINARY', 'XMLCAST', 'XMLCOMMENT', 'XMLCONCAT', 'XMLDECLARATION', 'XMLDOCUMENT', 'XMLELEMENT', 'XMLEXISTS', 'XMLFOREST', 'XMLITERATE', 'XMLNAMESPACES', 'XMLPARSE', 'XMLPI', 'XMLQUERY', 'XMLROOT', 'XMLSCHEMA', 'XMLSERIALIZE', 'XMLTABLE', 'XMLTEXT', 'XMLVALIDATE', 'YEAR', 'YES', 'ZONE']);
   function prepareSQLName(name) {
       if (isUnquotedIdentifier.test(name) && !reservedWords.has(name.toUpperCase()))
           return name;
   
       // Let's quote (and possibly escape) name:
       var reallyEscaped = false;
       var escapedName = '';
       for (var i = 0; i!=name.length; ++i) {
           var charCode = name.charCodeAt(i);
           if (charCode<128) {
               var char = name.charAt(i);
               escapedName += char=='"' ? '""' : char;
           }
           else {
               reallyEscaped = true;
               escapedName += '\\' + ('0000' + charCode.toString(16)).slice(-4);
           }
       }
       return reallyEscaped ? 'U&"' + escapedName + '"' : '"' + escapedName + '"';
   }
   
   console.log(prepareSQLName('simple')); //simple
   console.log(prepareSQLName('to "quote')); //"to ""quote"
   console.log(prepareSQLName('простий')); //U&"\043f\0440\043e\0441\0442\0438\0439"
   console.log(prepareSQLName('в "лапки')); //U&"\0432 ""\043b\0430\043f\043a\0438"
   

4. _{I recommend you to lowercase name by calling name = name.toLowerCase(); before quoting it. Without doing this you'll need to take care about the fact that hello and HeLLo would be treated as same name, but hello world and HeLLo WoRld as different (because the latter two will got quoted). But doing this will make your code incompatible with code that doesn't forcibly lowercase all names (i.e. you won't be able to access entities with non-all-lowercase names that contain something other than letter, digit, underscore or dollar sign).}

---

Details on detecting letters:

I'm not sure about this, but AFAIK different versions of Unicode standard can have different sets of code points for letters (i.e. some code point that is considered as letter in one version of Unicode standard may be considered as non-letter by other version of Unicode standard). If that's true, you'll get weird SQL syntax error, if JavaScript considers some code point as non-English letter and therefore sends it unquoted, while PostgreSQL considers it to be non-letter at all and therefore doesn't expect it outside quotes (e.g. due to older Unicode libraries within PostgreSQL server). To overcome this, you can either forcibly quote every non-ASCII character (by replacing \pL with A-Za-z), or at least ensure that JavaScript's definition of letter isn't wider than PostgreSQL (by replacing \pL with explicit code range list for specific Unicode version like here for Unicode 6.3.0).

De facto PostgreSQL allows almost any Unicode character (except for the forbidden characters within ASCII range and invalid surrogates) to be used within unquoted identifiers -- not just letters/digits/_/$ -- and even non-ASCII whitespace and unassigned code points are in fact allowed (although official documentation says nothing about that). This is probably done to prevent future compatibility issues (e.g. the one described in the strikeout text). The reasonable behavior at our side in this case would be to send only letters/digits/_/$ to PostgreSQL in unquoted form (as official documentation suggests), but now we don't need to worry about following exact version of Unicode standard for detecting them.

The question still remaining is whether we should treat only \pL as letters, or \pL\pM* too (the former may include letter with diacritic only if it's represented by single code point, while the latter includes all kinds of letters with diacritics). De jure official documentation says nothing for this question, while de facto PostgreSQL accepts almost any Unicode character. Probably, in such case the latter is more reasonable (that's why I replaced ^[\pL_][\pL_0-9$]*$ from original version of the first code listing with ^[\pL_][\pL\pM_0-9$]*$).

Resume:

1. Whether to escape (U&"\044F") or not to escape ("я") a character within quoted identifier -- it's absolutely the matter of your taste (per documentation, any character except for NUL may appear non-escaped in quoted identifier, but NUL isn't allowed within identifiers anyway).
2. Whether to quote ("hello") or not to quote (hello) an identifier:
   • an identifier must be quoted if it contains any ASCII character other than English letter, digit 0 to 9, underscore sign or dollar sign, or if it starts from any ASCII character other than English letter or underscore sign;
   • an identifier should be quoted if it contains any non-ASCII character other than letter of any language, with or without diacritics; de facto such identifier will be accepted even without quoting, but such behavior isn't mentioned in documentation and may be changed in future;
   • otherwise it's your choice whether to quote an identifier or no; if you prefer case insensitivity, you may tend to first listing (which does quoting at the minimum and no escaping), if you unlike non-ASCII characters, you may tend to second listing (which quotes and escapes every non-ASCII character, even officially allowed).

Answer 2

Given Bobby Tables won't be rearing his ugly head, you could try something like this, which expands upon klin's good attempt:

CREATE OR REPLACE FUNCTION prepareSQLName(VARCHAR(255), BOOLEAN)
RETURNS VARCHAR(255) LANGUAGE PLPGSQL AS $$
BEGIN
    IF ($2) THEN
        EXECUTE 'SELECT * FROM ' || $1;
    ELSE
        EXECUTE 'CREATE TEMP TABLE ' || $1 || ' (ID INT) ON COMMIT DROP';
    END IF;
    RETURN $1;
EXCEPTION WHEN OTHERS THEN
    IF LEFT($1, 1) <> '"' AND prepareSQLName('"' || $1 || '"', $2) = '"' || $1 || '"' THEN
        RETURN '"' || $1 || '"';
    ELSE
        RETURN NULL;
    END IF;
END $$;

Explanation

The question seems to include two cases:

1. Finding out whether an identifier is valid for an existing table ...and...
2. Finding out whether an identifier can be used for a new table or column name.

Hence the second boolean parameter in the function: This determines whether an attempt is made to select from an existing table (if true) or create a new temporary table (if false) - in both cases, first using the supplied identifier (first parameter value) as-is.

If no errors occurred the function simply returns the supplied identifier. If an error occurred, a second recursive call is made to see if wrapping in quotes helps the situation. If so, the quoted version is returned. But if this also caused an error then the identifier is deemed unusable in any form and NULL is returned.

See SQL Fiddle Demo for some example usage.

Additional Findings - Unicode table names

For the record, I wanted to find out how unicode behaved and which unicode characters are and aren't acceptable in the quoted form described in the link you posted. To do this, I tried similar selects to yours and found the Test3 table created in its quoted unicode form (CREATE TABLE U&"\0054\0065\0073\0074\0033" (ID INT);) behaved exactly the same as the quoted "Test2" table.

Also tried running this to find which characters can be used:

DO $do$
    DECLARE chr VARCHAR(9);
BEGIN
    FOR i IN 0..9999 LOOP
        chr := 'U&"\' || LPAD(CAST(i AS VARCHAR(4)), 4, '0') || '"';
        IF prepareSQLName(chr, FALSE) IS NULL THEN
            RAISE NOTICE 'NOT OK: %', chr;
        END IF;
    END LOOP;
END $do$;

It raises just one notice:

NOTICE:  NOT OK: U&"\0000"    

This shows the only unicode character that can't be used in a table name is the NUL character.