Working with Collations

1. Understanding Collation Options

Field	Values	Effect
locale	e.g. "en", "fr", "de_AT"	Language rules
strength	1-5	Comparison level (1=primary, 3=default)
caseLevel	bool	Case sensitivity at strength 1 or 2
caseFirst	"upper","lower","off"	Tiebreak order
numericOrdering	bool	"2" < "10" when true
alternate	"non-ignorable","shifted"	Punctuation handling
maxVariable	"punct","space"	Variable elements
backwards	bool	Reverse secondary (French)
normalization	bool	NFD normalize before compare

Form	Effect
`createCollection("c", {collation: {...}})`	Default for all ops
Immutable	Cannot change after creation

Form	Effect
`find(...).collation({...})`	Override per query
Index use	Requires matching index collation

db.users.find({ name: "alice" }).collation({ locale: "en", strength: 2 });

Form	Notes
`createIndex({name:1}, {collation:{...}})`	Indexed values sorted per collation
Match required	Query collation must equal index collation

strength	Effect
1	Base letters only
2	+ accents
3 (default)	+ case
caseLevel: true	Force case-significance at lower strengths

strength	café vs cafe
1	Equal
≥ 2	Different

Setting	Effect
numericOrdering: false (default)	"10" < "2" lexicographically
numericOrdering: true	"2" < "10" numerically

alternate	Effect
"non-ignorable" (default)	Punctuation participates
"shifted"	Spaces/punctuation ignored except at strength 4

Setting	Use
normalization: true	Compare canonical forms (NFD)
Cost	Slightly slower