Fix/table rows

[germa89]

Close #883.

[codecov]

Codecov Report

Merging #884 (3d2f534) into main (ebc75ac) will increase coverage by 0.07%.
The diff coverage is 94.11%.

@@            Coverage Diff             @@
##             main     #884      +/-   ##
==========================================
+ Coverage   59.23%   59.30%   +0.07%     
==========================================
  Files         172      172              
  Lines       10641    10641              
==========================================
+ Hits         6303     6311       +8     
+ Misses       4338     4330       -8     

[germa89]

I have realized that this fails:

>>> dimx = 10
>>> dimy = 1
>>> my_conv = np.random.rand(dimx, dimy)
>>> mapdl.load_array("my_conv", my_conv)
>>> assert np.allclose(mapdl.parameters["my_conv"], my_conv) #Fail
AssertionError                            Traceback (most recent call last)
<ipython-input-13-0db9fb3fa50d> in <module>
>>> mapdl.parameters["my_conv"].shape
(10,)
>>> my_conv.shape
(1, 10)

But this works

>>> dimx = 1
>>> dimy = 10
>>> my_conv = np.random.rand(dimx, dimy)
>>> mapdl.load_array("my_conv", my_conv)
>>> assert np.allclose(mapdl.parameters["my_conv"], my_conv) #Works
True
>>> mapdl.parameters["my_conv"].shape
(10,)
>>> my_conv.shape
(10, 1)

I could remove the squeeze of this line:

https://github.com/pyansys/pymapdl/blob/88624d499807f81f707372f9e71c9a42de0e7a80/ansys/mapdl/core/parameters.py#L368

However, then checking against a list will fail because:

>>> my_conv = [1,2,3]
>>> mapdl.load_array("my_conv", my_conv)
>>> assert np.allclose(mapdl.parameters["my_conv"], my_conv) #Works
False
>>> mapdl.parameters["my_conv"].shape
(3,1)
>>> my_conv.shape
(3,)

I guess we need to choose between the right assertions:
Either we return by default arrays of (X,1) or (X,).

If the former, list assertions will fail because they are initialized as (X,).
If the later, then the arrays assertion will fail because they are shaped as (X,1).

I feel we should go for the later, return always the second dimension even if it is 1.

[akaszynski]

I guess we need to choose between the right assertions:
Either we return by default arrays of (X,1) or (X,).

I remember writing this thinking that we might have this issue. Agree that we should be consistent, pick one and go with it (and I'm leaning (X,1).

I feel we should go for the later, return always the second dimension even if it is 1.

Implement and remind me to review.

[germa89]

@akaszynski ... feel free to review again.

[akaszynski]

LGTM.