Load_table results in table with one fewer rows than expected

[mikerife]

Describe the bug
load_table results in an MAPDL table with one fewer rows than expected. The last row of data from the inputted np.array is missing.

To Reproduce
Steps to reproduce the behavior:

1. Go to https://mapdldocs.pyansys.com/api/_autosummary/ansys.mapdl.core.Mapdl.load_table.html?highlight=load_table
2. Test example at bottom of page.

Expected behavior
Expected return is:
array([0.001, 0.001, 0.005, 0.005, 0.002, 0.002])

Screenshots

System Information:

• OS: Window 10 build 19043
• Ansys version 2021 R2 and 2022 R1
• PyMAPDL version 0.60.6

Run a PyMAPDL report

Please run the following code wherever you are experiencing the bug and paste the output below. This report helps us track down bugs and it is critical to addressing your bug:

from ansys.mapdl import core as pymapdl
print(pymapdl.Report())

# Paste your results here

-------------------------------------------------------------------------------
PyMAPDL Software and Environment Report
-------------------------------------------------------------------------------
  Date: Wed Feb 09 19:12:07 2022 Mountain Standard Time

                OS : Windows
            CPU(s) : 16
           Machine : AMD64
      Architecture : 64bit
       Environment : Python
       File system : unknown
NVIDIA Corporation : GPU Vendor
NVIDIA RTX A2000 Laptop GPU/PCIe/SSE2 : GPU Renderer
4.5.0 NVIDIA 472.88 : GPU Version

  Python 3.9.9 (tags/v3.9.9:ccb0e6a, Nov 15 2021, 18:08:50) [MSC v.1929 64 bit
  (AMD64)]

        matplotlib : 3.5.1
             numpy : 1.22.0
           pyvista : 0.33.2
           appdirs : 1.4.4
              tqdm : 4.62.3
            pyiges : 0.2.1
             scipy : 1.7.3
              grpc : 1.43.0
ansys.api.mapdl.v0 : 0.4.1
ansys.mapdl.reader : 0.51.8
   google.protobuf : 3.19.3
-------------------------------------------------------------------------------

Ansys Installation
******************
Version   Location
------------------
212       C:\Program Files\ANSYS Inc\v212
221       C:\Program Files\ANSYS Inc\v221


Ansys Environment Variables
***************************
ANSYS212_DIR                   C:\Program Files\ANSYS Inc\v212\ANSYS
ANSYS221_DIR                   C:\Program Files\ANSYS Inc\v221\ANSYS
ANSYS221_PRODUCT               ' '
ANSYSCLOUDCLI_ROOT             C:\Program Files\ANSYS Inc\ANSYS Cloud\ANSYS Cloud CLI\
ANSYSELECTRON_ROOT             C:\Program Files\ANSYS Inc\ANSYS Cloud\ANSYS Electron\
ANSYSLIC_DIR                   C:\Program Files\ANSYS Inc\Shared Files\Licensing
AWP_LOCALE212                  en-us
AWP_LOCALE221                  en-us
AWP_ROOT212                    C:\Program Files\ANSYS Inc\v212
AWP_ROOT221                    C:\Program Files\ANSYS Inc\v221
CADOE_LIBDIR212                C:\Program Files\ANSYS Inc\v212\CommonFiles\Language\en-us
CADOE_LIBDIR221                C:\Program Files\ANSYS Inc\v221\CommonFiles\Language\en-us

[mikerife]

I just realized too that the example has data of 0.001 etc but output is shown as 0.0001 etc.

[akaszynski]

@germa89, can you look at this?

[germa89]

Sure thing.

[germa89]

Ok. This is quite interesting.

Appart of the missing last row, I found out that, it seems that MAPDL have different behavior depending on how many columns you have in the table to be read.

Example

See the code attached.

>>> from ansys.mapdl.core import launch_mapdl
>>> import numpy as np
>>> mapdl = launch_mapdl()
>>> dimx = 3
>>> dimy = 3
>>> my_conv = np.random.rand(dimx, dimy)
>>> # Making headers in ascending order
>>> my_conv[:, 0] = np.arange(dimx)
>>> my_conv[0, :] = np.arange(dimy)*3+1
>>> print(my_conv)
[[1.         4.         7.        ]
 [1.         0.2942442  0.67930085]
 [2.         0.48780547 0.13856209]]

Array of 3 columns (or +2)

If you use the full 3x3:

>>> mapdl.load_table("my_conv", my_conv)
>>> print(mapdl.parameters["my_conv"])
[[0.2942442  0.67930085]
 [0.48780547 0.13856209]]
>>> print(mapdl.starstatus('my_conv')) # double check
PARAMETER STATUS- CONV  (      5 PARAMETERS DEFINED)
                  (INCLUDING        2 INTERNAL PARAMETERS)

      LOCATION                VALUE
        1       1       1   0.294244195
        2       1       1   0.487805469
        1       2       1   0.679300854
        2       2       1   0.138562089

Array of 2 columns

If you use the 3x2:

>>> my_conv2 = my_conv[:,:2]
>>> mapdl.load_table("my_conv2", my_conv2)
>>> print(mapdl.parameters["my_conv2"])
[4.         0.81696805]
>>> print(mapdl.starstatus('my_conv2')) # double check
PARAMETER STATUS- CONV2  (      5 PARAMETERS DEFINED)
                  (INCLUDING        2 INTERNAL PARAMETERS)

      LOCATION                VALUE
        1       1       1    4.00000000
        2       1       1   0.816968048

What is happening.

MAPDL (TREAD) when reading a 2 columns file, will NOT use the first row for the header.

However, if using more than 2 columns, it will use the first row for the header which is not later reported using STDOUT methods.

In the current implementation, there is no distinction, hence the mistake.

Not only that, np.squeeze is converting the column vector to a row vector.

Fix

It seems a MAPDL issue, therefore, we will just account for that case (array.size[1] ==2) and do the appropriate dimension arrangement. To be consistent, the header will not be reported in the case columns==2.