Identifiers

Regular identifiers accept Unicode letters and digits. Use delimited identifiers for spaces, punctuation or keyword-like names. Identifier spelling is preserved; catalog names are case-sensitive after decoding.

Query variables, property names and parameters compare without regard to ASCII case, the same way the parser rejects name and NAME as duplicates. The first spelling is the one stored, so SET n.AGE = 31 updates a property inserted as age. Labels stay case-sensitive: :Person and :person are different labels, and graph g is not graph G.

MATCH (n:Person) WHERE n.AGE >= 40
RETURN N.name AS name;

LET `display name` = 'Alice'
RETURN `display name` AS name;

The parser accepts both backtick and double-quoted delimited identifiers, escape sequences and no-escape prefixes. Character strings use single quotes. A dot in a delimited identifier belongs to the name; it is not a property-access separator.

Parameter names can be regular ($name), extended ($123) or delimited ($"tenant id"). The leading $ is GQL syntax; the Rust set_parameter API takes the decoded name without it.

SESSION SET VALUE $"tenant id" STRING = 'acme';
RETURN $"tenant id" AS tenant;

Catalog paths resolve directory/schema names separately from query binding variables.

Escapes and Unicode

An @ prefix disables backslash escaping in a delimited identifier. Unicode regular names need no delimiters when they satisfy identifier rules.

LET 名字 = 'Alice', @`raw\nname` = 'Bob'
RETURN 名字 AS name, @`raw\nname` AS raw_name;

Here the raw identifier contains the literal characters backslash and n, not a newline.