Data Types in NoSQL (MongoDB)

Technical document on data types in MongoDB, their internal structure, and recommended applications, with a practical focus for flexible modeling, efficient querying, and compatibility with distributed systems.

---

1. Primitive Types

Type	BSON Equivalent	Description	Recommendations for Use
String	String	UTF-8 text string	Names, addresses, textual content
NumberInt	Int32	32-bit integer	Counters, flags, small ranges
NumberLong	Int64	64-bit integer	Timestamps, large accumulators
NumberDouble	Double	Double precision floating point	Statistical calculations, approximate values
Boolean	Boolean	True or false	Logical states, enable/disable
Date	Date	Date and time in milliseconds since epoch	Time logging, TTL
Null	Null	Absence of value	Optional fields, semantic consistency

---

2. Complex Types

Type	BSON	Description	Recommendations for Use
Object	Document	Nested object with subfields	Modeling hierarchical entities
Array	Array	Ordered list of elements	1:N relationships, value series
Embedded Document	Document	Embedded subdocument	Avoid JOINs, improve read efficiency
ObjectId	ObjectId	12-byte hexadecimal unique ID	Primary keys, document references
Binary	Binary	Binary data (Buffer)	Files, images, encoded tokens
Code	JavaScript	Executable JS code snippets	Advanced validations or filters (limited use)

---

3. Special and System Types

Type	BSON	Description	Recommendations for Use
Timestamp	Timestamp	Timestamp with insertion order	Sequential logs, internal replication
Decimal128	Decimal128	High precision decimal (128 bits)	Finance, exact values
MinKey	MinKey	Lowest possible value	Lower limit comparisons
MaxKey	MaxKey	Highest possible value	Upper limit comparisons
Undefined	Undefined	Undefined value (deprecated)	Not recommended, incompatible in new versions

---

4. General Recommendations

• Flexible by design: NoSQL allows variable structure, ideal for heterogeneous or evolving data.
• Embedded documents: Improve read performance; avoid over-nesting that makes updates difficult.
• References vs. Embedding: Use references (ObjectId) if data is shared across collections.
• Decimal128: Preferred in financial systems; avoid using Double in monetary calculations.
• Array Indexing: MongoDB allows indexing individual elements of an array, useful for complex filtering.
• ⏱ TTL and dates: Combine Date fields with TTL Index to automatically delete documents.
• Avoid deprecated types: Undefined and Code are rarely used and have security and compatibility implications.

---