Spark SQL Data Types: Primitives, Nested, and struct vs map

The set of Spark types you’ll actually run into in real projects:

string
integer
array
struct
long
boolean
double
map
binary
float
decimal(20,0)
timestamp
date

Common types

Type	Spark SQL name	Description
string	STRING / StringType	variable-length Unicode text
integer	INT / IntegerType	4-byte signed integer
long	BIGINT / LongType	8-byte signed integer
boolean	BOOLEAN / BooleanType	true / false
double	DOUBLE / DoubleType	8-byte IEEE 754 double-precision float
float	FLOAT / FloatType	4-byte single-precision float
decimal(20,0)	DECIMAL(20,0)	20-digit exact number, 0 decimal places (i.e. an exact 20-digit integer), good for financial/precise values
timestamp	TIMESTAMP / TimestampType	date + time, microsecond precision, usually UTC
date	DATE / DateType	year-month-day only
binary	BINARY / BinaryType	variable-length byte array (raw binary)

Nested types

Type	Spark SQL name	Description
array	ARRAY<T>	ordered list, all elements the same type T
map	MAP<K,V>	key-value pairs; all keys share a type, all values share a type
struct	STRUCT<...>	a composite record with named fields, each field with its own type; each key-value pair can have its own type; more flexible than map (a subset)

---

References

• Spark SQL Data Types: https://spark.apache.org/docs/latest/sql-ref-datatypes.html

Related: back to the Cross-Engine DB Query overview, or on to Spark array functions and collect.