pyarrow.Array — Apache Arrow v20.0.0 (original) (raw)
class pyarrow.Array#
Bases: _PandasConvertible
The base class for all Arrow arrays.
__init__(*args, **kwargs)#
Methods
Attributes
buffers(self)#
Return a list of Buffer objects pointing to this array’s physical storage.
To correctly interpret these buffers, you need to also apply the offset multiplied with the size of the stored data type.
cast(self, target_type=None, safe=None, options=None, memory_pool=None)#
Cast array values to another data type
See pyarrow.compute.cast() for usage.
Parameters:
target_typeDataType, default None
Type to cast array to.
Whether to check for conversion errors such as overflow.
optionsCastOptions, default None
Additional checks pass by CastOptions
memory_poolMemoryPool, optional
memory pool to use for allocations during function execution.
Returns:
castArray
copy_to(self, destination)#
Construct a copy of the array with all buffers on destination device.
This method recursively copies the array’s buffers and those of its children onto the destination MemoryManager device and returns the new Array.
Parameters:
destinationpyarrow.MemoryManager or pyarrow.Device
The destination device to copy the array to.
Returns:
device_type#
The device type where the array resides.
Returns:
DeviceAllocationType
dictionary_encode(self, null_encoding='mask')#
Compute dictionary-encoded representation of array.
See pyarrow.compute.dictionary_encode() for full usage.
Parameters:
null_encodingstr, default “mask”
How to handle null entries.
Returns:
encodedDictionaryArray
A dictionary-encoded version of this array.
diff(self, Array other)#
Compare contents of this array against another one.
Return a string containing the result of diffing this array (on the left side) against the other array (on the right side).
Parameters:
otherArray
The other array to compare this array with.
Returns:
diffstr
A human-readable printout of the differences.
Examples
import pyarrow as pa left = pa.array(["one", "two", "three"]) right = pa.array(["two", None, "two-and-a-half", "three"]) print(left.diff(right))
@@ -0, +0 @@ -“one” @@ -2, +1 @@ +null +”two-and-a-half”
drop_null(self)#
Remove missing values from an array.
equals(self, Array other)#
Parameters:
otherpyarrow.Array
Returns:
fill_null(self, fill_value)#
See pyarrow.compute.fill_null() for usage.
Parameters:
fill_valueany
The replacement value for null entries.
Returns:
resultArray
A new array with nulls replaced by the given value.
filter(self, mask, *, null_selection_behavior='drop')#
Select values from an array.
See pyarrow.compute.filter() for full usage.
Parameters:
maskArray or array-like
The boolean mask to filter the array with.
null_selection_behaviorstr, default “drop”
How nulls in the mask should be handled.
Returns:
filteredArray
An array of the same type, with only the elements selected by the boolean mask.
format(self, **kwargs)#
DEPRECATED, use pyarrow.Array.to_string
Parameters:
**kwargsdict
Returns:
static from_buffers(DataType type, length, buffers, null_count=-1, offset=0, children=None)#
Construct an Array from a sequence of buffers.
The concrete type returned depends on the datatype.
Parameters:
typeDataType
The value type of the array.
lengthint
The number of values in the array.
buffersList[Buffer]
The buffers backing this array.
null_countint, default -1
The number of null entries in the array. Negative value means that the null count is not known.
offsetint, default 0
The array’s logical offset (in values, not in bytes) from the start of each buffer.
childrenList[Array], default None
Nested type children with length matching type.num_fields.
Returns:
arrayArray
static from_pandas(obj, mask=None, type=None, bool safe=True, MemoryPool memory_pool=None)#
Convert pandas.Series to an Arrow Array.
This method uses Pandas semantics about what values indicate nulls. See pyarrow.array for more general conversion from arrays or sequences to Arrow arrays.
Parameters:
objndarray, pandas.Series, array-like
Indicate which values are null (True) or not null (False).
typepyarrow.DataType
Explicit type to attempt to coerce to, otherwise will be inferred from the data.
Check for overflows or other unsafe conversions.
memory_poolpyarrow.MemoryPool, optional
If not passed, will allocate memory from the currently-set default memory pool.
Returns:
arraypyarrow.Array or pyarrow.ChunkedArray
ChunkedArray is returned if object data overflows binary buffer.
Notes
Localized timestamps will currently be returned as UTC (pandas’s native representation). Timezone-naive data will be implicitly interpreted as UTC.
get_total_buffer_size(self)#
The sum of bytes in each buffer referenced by the array.
An array may only reference a portion of a buffer. This method will overestimate in this case and return the byte size of the entire buffer.
If a buffer is referenced multiple times then it will only be counted once.
index(self, value, start=None, end=None, *, memory_pool=None)#
Find the first index of a value.
See pyarrow.compute.index() for full usage.
Parameters:
valueScalar or object
The value to look for in the array.
startint, optional
The start index where to look for value.
endint, optional
The end index where to look for value.
memory_poolMemoryPool, optional
A memory pool for potential memory allocations.
Returns:
indexInt64Scalar
The index of the value in the array (-1 if not found).
is_cpu#
Whether the array is CPU-accessible.
is_nan(self)#
Return BooleanArray indicating the NaN values.
Returns:
is_null(self, *, nan_is_null=False)#
Return BooleanArray indicating the null values.
Parameters:
nan_is_nullbool (optional, default False)
Whether floating-point NaN values should also be considered null.
Returns:
is_valid(self)#
Return BooleanArray indicating the non-null values.
nbytes#
Total number of bytes consumed by the elements of the array.
In other words, the sum of bytes from all buffer ranges referenced.
Unlike get_total_buffer_size this method will account for array offsets.
If buffers are shared between arrays then the shared portion will be counted multiple times.
The dictionary of dictionary arrays will always be counted in their entirety even if the array only references a portion of the dictionary.
null_count#
offset#
A relative position into another array’s data.
The purpose is to enable zero-copy slicing. This value defaults to zero but must be applied on all operations with the physical storage buffers.
slice(self, offset=0, length=None)#
Compute zero-copy slice of this array.
Parameters:
offsetint, default 0
Offset from start of array to slice.
Length of slice (default is until end of Array starting from offset).
Returns:
slicedArray
An array with the same datatype, containing the sliced values.
sort(self, order='ascending', **kwargs)#
Sort the Array
Parameters:
orderstr, default “ascending”
Which order to sort values in. Accepted values are “ascending”, “descending”.
**kwargsdict, optional
Additional sorting options. As allowed by SortOptions
Returns:
resultArray
statistics#
Statistics of the array.
sum(self, **kwargs)#
Sum the values in a numerical array.
See pyarrow.compute.sum() for full usage.
Parameters:
**kwargsdict, optional
Options to pass to pyarrow.compute.sum().
Returns:
sumScalar
A scalar containing the sum value.
take(self, indices)#
Select values from an array.
See pyarrow.compute.take() for full usage.
Parameters:
indicesArray or array-like
The indices in the array whose values will be returned.
Returns:
takenArray
An array with the same datatype, containing the taken values.
to_numpy(self, zero_copy_only=True, writable=False)#
Return a NumPy view or copy of this array.
By default, tries to return a view of this array. This is only supported for primitive arrays with the same memory layout as NumPy (i.e. integers, floating point, ..) and without any nulls.
For the extension arrays, this method simply delegates to the underlying storage array.
Parameters:
zero_copy_onlybool, default True
If True, an exception will be raised if the conversion to a numpy array would require copying the underlying data (e.g. in presence of nulls, or for non-primitive types).
For numpy arrays created with zero copy (view on the Arrow data), the resulting array is not writable (Arrow data is immutable). By setting this to True, a copy of the array is made to ensure it is writable.
Returns:
arraynumpy.ndarray
to_pandas(self, memory_pool=None, categories=None, bool strings_to_categorical=False, bool zero_copy_only=False, bool integer_object_nulls=False, bool date_as_object=True, bool timestamp_as_object=False, bool use_threads=True, bool deduplicate_objects=True, bool ignore_metadata=False, bool safe=True, bool split_blocks=False, bool self_destruct=False, unicode maps_as_pydicts=None, types_mapper=None, bool coerce_temporal_nanoseconds=False)#
Convert to a pandas-compatible NumPy array or DataFrame, as appropriate
Parameters:
memory_poolMemoryPool, default None
Arrow MemoryPool to use for allocations. Uses the default memory pool if not passed.
categorieslist, default empty
List of fields that should be returned as pandas.Categorical. Only applies to table-like data structures.
strings_to_categoricalbool, default False
Encode string (UTF8) and binary types to pandas.Categorical.
zero_copy_onlybool, default False
Raise an ArrowException if this function call would require copying the underlying data.
integer_object_nullsbool, default False
Cast integers with nulls to objects
date_as_objectbool, default True
Cast dates to objects. If False, convert to datetime64 dtype with the equivalent time unit (if supported). Note: in pandas version < 2.0, only datetime64[ns] conversion is supported.
timestamp_as_objectbool, default False
Cast non-nanosecond timestamps (np.datetime64) to objects. This is useful in pandas version 1.x if you have timestamps that don’t fit in the normal date range of nanosecond timestamps (1678 CE-2262 CE). Non-nanosecond timestamps are supported in pandas version 2.0. If False, all timestamps are converted to datetime64 dtype.
Whether to parallelize the conversion using multiple threads.
deduplicate_objectsbool, default True
Do not create multiple copies Python objects when created, to save on memory use. Conversion will be slower.
ignore_metadatabool, default False
If True, do not use the ‘pandas’ metadata to reconstruct the DataFrame index, if present
For certain data types, a cast is needed in order to store the data in a pandas DataFrame or Series (e.g. timestamps are always stored as nanoseconds in pandas). This option controls whether it is a safe cast or not.
split_blocksbool, default False
If True, generate one internal “block” for each column when creating a pandas.DataFrame from a RecordBatch or Table. While this can temporarily reduce memory note that various pandas operations can trigger “consolidation” which may balloon memory use.
self_destructbool, default False
EXPERIMENTAL: If True, attempt to deallocate the originating Arrow memory while converting the Arrow object to pandas. If you use the object after calling to_pandas with this option it will crash your program.
Note that you may not see always memory usage improvements. For example, if multiple columns share an underlying allocation, memory can’t be freed until all columns are converted.
maps_as_pydictsstr, optional, default None
Valid values are None, ‘lossy’, or ‘strict’. The default behavior (None), is to convert Arrow Map arrays to Python association lists (list-of-tuples) in the same order as the Arrow Map, as in [(key1, value1), (key2, value2), …].
If ‘lossy’ or ‘strict’, convert Arrow Map arrays to native Python dicts. This can change the ordering of (key, value) pairs, and will deduplicate multiple keys, resulting in a possible loss of data.
If ‘lossy’, this key deduplication results in a warning printed when detected. If ‘strict’, this instead results in an exception being raised when detected.
types_mapperfunction, default None
A function mapping a pyarrow DataType to a pandas ExtensionDtype. This can be used to override the default pandas type for conversion of built-in pyarrow types or in absence of pandas_metadata in the Table schema. The function receives a pyarrow DataType and is expected to return a pandas ExtensionDtype or None if the default conversion should be used for that type. If you have a dictionary mapping, you can pass dict.get as function.
coerce_temporal_nanosecondsbool, default False
Only applicable to pandas version >= 2.0. A legacy option to coerce date32, date64, duration, and timestamp time units to nanoseconds when converting to pandas. This is the default behavior in pandas version 1.x. Set this option to True if you’d like to use this coercion when using pandas version >= 2.0 for backwards compatibility (not recommended otherwise).
Returns:
pandas.Series or pandas.DataFrame depending on type of object
Examples
import pyarrow as pa import pandas as pd
Convert a Table to pandas DataFrame:
table = pa.table([ ... pa.array([2, 4, 5, 100]), ... pa.array(["Flamingo", "Horse", "Brittle stars", "Centipede"]) ... ], names=['n_legs', 'animals']) table.to_pandas() n_legs animals 0 2 Flamingo 1 4 Horse 2 5 Brittle stars 3 100 Centipede isinstance(table.to_pandas(), pd.DataFrame) True
Convert a RecordBatch to pandas DataFrame:
import pyarrow as pa n_legs = pa.array([2, 4, 5, 100]) animals = pa.array(["Flamingo", "Horse", "Brittle stars", "Centipede"]) batch = pa.record_batch([n_legs, animals], ... names=["n_legs", "animals"]) batch pyarrow.RecordBatch n_legs: int64 animals: string
n_legs: [2,4,5,100] animals: ["Flamingo","Horse","Brittle stars","Centipede"]
batch.to_pandas() n_legs animals 0 2 Flamingo 1 4 Horse 2 5 Brittle stars 3 100 Centipede isinstance(batch.to_pandas(), pd.DataFrame) True
Convert a Chunked Array to pandas Series:
import pyarrow as pa n_legs = pa.chunked_array([[2, 2, 4], [4, 5, 100]]) n_legs.to_pandas() 0 2 1 2 2 4 3 4 4 5 5 100 dtype: int64 isinstance(n_legs.to_pandas(), pd.Series) True
to_pylist(self, *, maps_as_pydicts=None)#
Convert to a list of native Python objects.
Parameters:
maps_as_pydictsstr, optional, default None
Valid values are None, ‘lossy’, or ‘strict’. The default behavior (None), is to convert Arrow Map arrays to Python association lists (list-of-tuples) in the same order as the Arrow Map, as in [(key1, value1), (key2, value2), …].
If ‘lossy’ or ‘strict’, convert Arrow Map arrays to native Python dicts.
If ‘lossy’, whenever duplicate keys are detected, a warning will be printed. The last seen value of a duplicate key will be in the Python dictionary. If ‘strict’, this instead results in an exception being raised when detected.
Returns:
lstlist
to_string(self, *, int indent=2, int top_level_indent=0, int window=10, int container_window=2, bool skip_new_lines=False)#
Render a “pretty-printed” string representation of the Array.
Note: for data on a non-CPU device, the full array is copied to CPU memory.
Parameters:
indentint, default 2
How much to indent the internal items in the string to the right, by default 2.
top_level_indentint, default 0
How much to indent right the entire content of the array, by default 0.
windowint
How many primitive items to preview at the begin and end of the array when the array is bigger than the window. The other items will be ellipsed.
container_windowint
How many container items (such as a list in a list array) to preview at the begin and end of the array when the array is bigger than the window.
skip_new_linesbool
If the array should be rendered as a single line of text or if each element should be on its own line.
tolist(self)#
Alias of to_pylist for compatibility with NumPy.
type#
unique(self)#
Compute distinct elements in array.
Returns:
uniqueArray
An array of the same data type, with deduplicated elements.
validate(self, *, full=False)#
Perform validation checks. An exception is raised if validation fails.
By default only cheap validation checks are run. Pass full=Truefor thorough validation checks (potentially O(n)).
Parameters:
If True, run expensive checks, otherwise cheap checks only.
Raises:
ArrowInvalid
value_counts(self)#
Compute counts of unique elements in array.
Returns:
An array of <input type “Values”, int64 “Counts”> structs
view(self, target_type)#
Return zero-copy “view” of array as another data type.
The data types must have compatible columnar buffer layouts
Parameters:
target_typeDataType
Type to construct view as.
Returns:
viewArray