Pandas Array/Scalar/Data Type Reference Manual

For most data types, Pandas uses NumPy arrays as the concrete storage objects, which are contained in Index, Series, or DataFrame.

For certain data types, Pandas extends NumPy's type system. String aliases for these types can be found in dtypes.

Pandas Array

Class/Method Description
pd.array(data, dtype) Create a Pandas array (ExtensionArray）。
pd.Series.array Return the underlying array of a Series (ExtensionArray）。
pd.arrays.IntegerArray Array for storing integer data (supports missing values).
pd.arrays.BooleanArray Array for storing boolean data (supports missing values).
pd.arrays.StringArray Array for storing string data (supports missing values).
pd.arrays.IntervalArray Array for storing interval data.
pd.arrays.DatetimeArray Array for storing datetime data.
pd.arrays.TimedeltaArray Array for storing timedelta data.
pd.arrays.PeriodArray Array for storing period data.
pd.arrays.SparseArray Array for storing sparse data.

Pandas Scalar

Class/Method Description
pd.NA Scalar representing missing values (similar toNaN）。
pd.Timestamp Scalar representing timestamps.
pd.Timedelta Scalar representing time differences.
pd.Period Scalar representing periods.
pd.Interval Scalar representing intervals.
pd.Categorical Scalar representing categorical data.

Pandas Data Types

Class/Method Description
pd.StringDtype() String data type (supports missing values).
pd.BooleanDtype() Boolean data type (supports missing values).
pd.Int8Dtype() 8-bit integer data type (supports missing values).
pd.Int16Dtype() 16-bit integer data type (supports missing values).
pd.Int32Dtype() 32-bit integer data type (supports missing values).
pd.Int64Dtype() 64-bit integer data type (supports missing values).
pd.Float32Dtype() 32-bit float data type (supports missing values).
pd.Float64Dtype() 64-bit float data type (supports missing values).
pd.CategoricalDtype() Categorical data type.
pd.DatetimeTZDtype() Datetime data type with timezone.
pd.PeriodDtype() Period data type.
pd.IntervalDtype() Interval data type.
pd.SparseDtype() Sparse data type.

Common Methods

Array Methods

Method Description
array.take(indices) Extract elements from an array based on indices.
array.copy() Copy the array.
array.isna() Check for missing values in the array.
array.fillna(value) Fill missing values with a specified value.
array.unique() Return unique values in the array.
array.value_counts() Return the frequency of each value in the array.

Scalar Methods

Method Description
timestamp.to_pydatetime() willTimestampConvert to Python'sdatetimeobject.
timedelta.total_seconds() willTimedeltaConvert to total seconds.
period.start_time ReturnPeriodthe start time of
period.end_time ReturnPeriodthe end time of
interval.left ReturnIntervalthe left boundary of
interval.right ReturnIntervalthe right boundary of

Data Type Methods

Method Description
dtype.name Return the name of the data type.
dtype.kind Return the kind of the data type (e.g.,imeans integer,fmeans float).
dtype.construct_array_type() Return the array class associated with the data type.

Examples

import pandas as pd

# Create a Pandas array
arr = pd.array([1, 2, None], dtype=pd.Int64Dtype())
print(arr)

# Use Pandas scalar
ts = pd.Timestamp('2023-01-01')
print(ts.year)  # Output the year

# Use Pandas data type
dtype = pd.StringDtype()
print(dtype.name)  # Output data type name

For more detailed information, please refer toPandas official documentation。

Other Extensions