Home "No matching signature found" error in python scipy ndimage.label() function used with boolean array

Questions

"No matching signature found" error in python scipy ndimage.label() function used with boolean array

November 9, 2023

I have a boolean array to indicate dry days:

dry_bool

0        False
1        False
2        False
3        False
4        False
...  
15336    False
15337    False
15338    False
15339    False
15340    False
Name: budget, Length: 15341, dtype: object

The number of False and True values are:

dry_bool.value_counts()

False    14594
True       747
Name: budget, dtype: int64

I want to find contiguous periods of dry days (where dry_bool=True).
Using the ndimage.label() function by the scipy package

import scipy.ndimage as ndimage

events, n_events = ndimage.label(dry_bool)

gives me the following error:

---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
Cell In[76], line 2
      1 # Find contiguous regions of dry_bool = True
----> 2 events, n_events = ndimage.label(dry_bool)

File ~/opt/anaconda3/lib/python3.9/site-packages/scipy/ndimage/_measurements.py:219, in label(input, structure, output)
    216         return output, maxlabel
    218 try:
--> 219     max_label = _ni_label._label(input, structure, output)
    220 except _ni_label.NeedMoreBits as e:
    221     # Make another attempt with enough bits, then try to cast to the
    222     # new type.
    223     tmp_output = np.empty(input.shape, np.intp if need_64bits else np.int32)

File _ni_label.pyx:202, in _ni_label._label()

File _ni_label.pyx:239, in _ni_label._label()

File _ni_label.pyx:95, in _ni_label.__pyx_fused_cpdef()

TypeError: No matching signature found

which I cannot wrap my head around. Any clue what’s going on?

There are definitely consecutive dry days, as you can see when I read the indices of the True values in the dry_bool array:

dry_idcs = [i for i, x in enumerate(dry_bool) if x]
print(dry_idcs[0:10])

[165, 205, 206, 214, 229, 230, 262, 281, 292, 301]

>Solution :

In your first code snippet it says dtype: object, even though it’s supposed to be an array of data type boolean (not object). Given that your error message gives you a TypeError, maybe that’s the problem?

You could try explicitly converting the data type of your array: