Closed
Description
Even after the fixes in pull request #6094, there are still some cases where string representation fails, in particular where some members of a structured array are multi-dimensional:
In [1]: ar = array((0, [[0.0, 0.0, 0.0], [0.0, 0.0, 0.0]], 0.0), dtype="int, (2,3)float, float")
In [2]: arm = ma.masked_array(ar)
In [3]: x = str(arm)
---------------------------------------------------------------------------
ValueError Traceback (most recent call last)
<ipython-input-3-f98bbc78fe11> in <module>()
----> 1 x = str(arm)
/home/users/gholl/venv/stable-3.5/lib/python3.5/site-packages/numpy/ma/core.py in __str__(self)
3698 if m.shape == ():
3699 if m.dtype.names:
-> 3700 m = m.view((bool, len(m.dtype)))
3701 if m.any():
3702 return str(tuple((f if _m else _d) for _d, _m in
ValueError: new type not compatible with array.
And without the m = m.view(...) line, m.any() would fail with TypeError: cannot perform reduce with flexible type.
A bit more of an overview, illustrating what we would like to see:
In [26]: m1 = ma.masked_array(data=[(0, [0, 0], 0)], mask=[(True,[False, True],True)], dtype="int, (2)float, int")
In [27]: m2 = ma.masked_array(data=(0, [0, 0], 0), mask=(True,[False, True],True), dtype="int, (2)float, int")
In [28]: str(m1)
Out[28]: '[(--, [0.0, --], --)]'
In [29]: str(m1[0])
Out[29]: '(--, [0.0, --], --)'
In [30]: str(m2)
---------------------------------------------------------------------------
ValueError Traceback (most recent call last)
<ipython-input-30-3ae8d0bbdb8a> in <module>()
----> 1 str(m2)
/home/users/gholl/venv/stable-3.5/lib/python3.5/site-packages/numpy/ma/core.py in __str__(self)
3698 if m.shape == ():
3699 if m.dtype.names:
-> 3700 m = m.view((bool, len(m.dtype)))
3701 if m.any():
3702 return str(tuple((f if _m else _d) for _d, _m in
ValueError: new type not compatible with array.
In [34]: numpy.ma.core.mvoid.__str__(m2)
Out[34]: '(--, [0.0, --], --)'