Data Types

pandas uses NumPy arrays and dtypes for Series or individual columns of a DataFrame. pandas and third-party libraries extend NumPy’s type system in a few places.

By default integer types are int64, float types are float64 and string types are object.

Data Type Info

Use the dtypes attribute to display the data type of each column in a DataFrame.

Change Data Types

When reading in data the default types of object, float64 and int64 are applied to columns. The column data types can be specified in various ways.

Memory Usage

Changes the data types of columns of a DataFrame can drastically reduce its memory usage.

Date

pandas has created a tremendous amount of functionality for manipulating time series data.

Date time

Date times are specific dates and times with timezone support.

Time delta

Time deltas are an absolute time duration.

Read Date Column Performance

Compare different ways of converting to timestamps.

pandas Data Types

The following table contains some of the pandas data types with some notes on each type.

String Alias	Note	Size	Alias	Types
bool	True or False	8		bool, np.bool
int64	[-9,223,372,036,854,775,808, 9,223,372,036,854,775,807). Default for integers.	64	long	int, np.int64
float64	Double precision float: sign bit, 11 bits exponent, 52 bits mantissa. Default for	64	double	float, np.float64
object	Can hold any Python object. Default for string data.		str	object, str, np.object_
string	Stores text data.			str, pd.StringDtype
timedelta64[ns]	Represents a duration, the difference between two dates or times.			datetime.timedelta, np.timedelta64, pd.Timedelta
datetime64[ns]	A specific date and time with timezone support.			datetime.datetime, np.datetime64, pd.DatetimeTZDtype
category	Useful for object columns with relatively few unique values.			pd.Categorical

int8	[-128, 128]	8	byte	np.int8
uint8	[0, 255]	8	ubyte	np.uint8
int16	[-32,768, 32,767]	16	short	np.int16
uint16	[0, 65,535]	16	ushort	np.uint16
int32	[-2,147,483,648, 2,147,483,647]	32		np.int32
uint32	[0, 4,294,967,295]	32		np.uint32
uint64	[0, +18,446,744,073,709,551,615]	64	uint	np.uint64
float16	Half precision float: sign bit, 5 bits exponent, 10 bits mantissa	16	half	np.float16
float32	Single precision float: sign bit, 8 bits exponent, 23 bits mantissa	32	single	np.float32
float128	Extended precision float: sign bit, 15 bits exponent, 112 bits mantissa.	128	longfloat	np.float128
complex64	Two 32-bit floats (real and imaginary components).	64	csingle	np.complex64
complex128	Two 64-bit floats (real and imaginary components).	128	cfloat	complex, np.complex128
complex256	Two 128-bit floats (real and imaginary components).	256	clongfloat	np.complex256
bytes	An array of bytes.			np.bytes_