On 2010-07-11, Thomas Jollans <tho...@jollans.com> wrote:
> On 07/11/2010 08:45 PM, wheres pythonmonks wrote:
>> On #3: Sorry this is confusing, but I was browsing some struct array
>> code from numpy, in which one of the columns contained strings, but
>> the type information, supplied in numpy.array's dtype argument,
>> specified the type as a an "object" not a string.
A string is an object.
>> Just wondering why one would do that.
>
> No expert on numpy, but maybe storing object references is cheaper than
> storing strings here ?
Strings are objects.
IIRC, numpy has special homogeneous array types to hold certain scalar
values: byte, int, float, complex. Those special array types allow
for very efficient storage and operations.
If you want an array of any other type, or a heterogeneous array, then
you use an array of objects and then you can put anything into the
array (including strings).
--
Grant Edwards grant.b.edwards Yow! The PINK SOCKS were
at ORIGINALLY from 1952!!
gmail.com But they went to MARS
around 1953!!
--
http://mail.python.org/mailman/listinfo/python-list
