ansys.dyna.core.lib.array
=========================

.. py:module:: ansys.dyna.core.lib.array


Functions
---------

.. autoapisummary::

   ansys.dyna.core.lib.array.array


Module Contents
---------------

.. py:function:: array(element_type: type, reserved_size: int = 0, default_value=None)

   
   A resizable array that supports optional values for any type
   Right now - array.array is used for single and double precision floating points
   for everything else - python list is used.  This is because no existing array
   type in numpy, pandas, and python meet the above requirements.  Specifically,
   numpy integer arrays do not have optional values and are not resizable, pandas
   integer arrays support optional values but are also not resizable, while python
   array arrays are resizable but do not support optional values.

   The problem with this approach is memory usage.  For 100k integers, a python list
   appears to take about 5300K, while a pandas array and numpy array take 488K and 584K
   respectively.  pandas arrays take more memory than numpy because of the masking used
   to support optional integer values.

   Given a python list of optional integer, where None is used to represent a missing value,
   - this is how you convert to either type:
   numpy:  np.array([item or 0 for item in the_list], dtype=np.int32)
   pandas: pd.array(the_list,dtype=pd.Int32Dtype())

   In the future - A dynamic array class based on some of the above types can be used for
   integer arrays.  For string arrays, pandas arrays don't offer any value over python
   lists.















   ..
       !! processed by numpydoc !!