MCPcopy Create free account
hub / github.com/apache/arrow / cast

Function cast

python/pyarrow/compute.py:349–415  ·  view source on GitHub ↗

Cast array values to another data type. Can also be invoked as an array instance method. Parameters ---------- arr : Array-like target_type : DataType or str Type to cast to safe : bool, default True Check for overflows or other unsafe conversions op

(arr, target_type=None, safe=None, options=None, memory_pool=None)

Source from the content-addressed store, hash-verified

347
348
349def cast(arr, target_type=None, safe=None, options=None, memory_pool=None):
350 """
351 Cast array values to another data type. Can also be invoked as an array
352 instance method.
353
354 Parameters
355 ----------
356 arr : Array-like
357 target_type : DataType or str
358 Type to cast to
359 safe : bool, default True
360 Check for overflows or other unsafe conversions
361 options : CastOptions, default None
362 Additional checks pass by CastOptions
363 memory_pool : MemoryPool, optional
364 memory pool to use for allocations during function execution.
365
366 Examples
367 --------
368 >>> from datetime import datetime
369 >>> import pyarrow as pa
370 >>> arr = pa.array([datetime(2010, 1, 1), datetime(2015, 1, 1)])
371 >>> arr.type
372 TimestampType(timestamp[us])
373
374 You can use ``pyarrow.DataType`` objects to specify the target type:
375
376 >>> cast(arr, pa.timestamp('ms'))
377 <pyarrow.lib.TimestampArray object at ...>
378 [
379 2010-01-01 00:00:00.000,
380 2015-01-01 00:00:00.000
381 ]
382
383 >>> cast(arr, pa.timestamp('ms')).type
384 TimestampType(timestamp[ms])
385
386 Alternatively, it is also supported to use the string aliases for these
387 types:
388
389 >>> arr.cast('timestamp[ms]')
390 <pyarrow.lib.TimestampArray object at ...>
391 [
392 2010-01-01 00:00:00.000,
393 2015-01-01 00:00:00.000
394 ]
395 >>> arr.cast('timestamp[ms]').type
396 TimestampType(timestamp[ms])
397
398 Returns
399 -------
400 casted : Array
401 The cast result as a new Array
402 """
403 safe_vars_passed = (safe is not None) or (target_type is not None)
404
405 if safe_vars_passed and (options is not None):
406 raise ValueError("Must either pass values for 'target_type' and 'safe'"

Callers

nothing calls this directly

Calls

no outgoing calls

Tested by

no test coverage detected