I have two Numpy record arrays that have exactly the same fields. What is the easiest way to combine them into one (i.e. append one table on to the other)?
A:
for i in array1:
array2.append(i)
Or (if implemented)
array1.extend(array2)
Now array1 contains also all elements of array2
Lex
2009-11-10 15:28:07
A:
#!/usr/bin/env python
import numpy as np
desc = {'names': ('gender','age','weight'), 'formats': ('S1', 'f4', 'f4')}
a = np.array([('M',64.0,75.0),('F',25.0,60.0)], dtype=desc)
b = np.array([('M',64.0,75.0),('F',25.0,60.0)], dtype=desc)
alen=a.shape[0]
blen=b.shape[0]
a.resize(alen+blen)
a[alen:]=b[:]
This works with structured arrays, though not recarrays. Perhaps this is a good reason to stick with structured arrays.
unutbu
2009-11-10 15:37:47
Is there a reason why this does not work with recarrays? I thought recarrays were just structured arrays with an extra __getattribute__/__setattr__ arguments?
astrofrog
2009-11-10 15:55:43
I don't know why. I only know that when I try the same thing with recarrays I get a ValueError: cannot resize this array: it does not own its own data. Having run into problems like this with recarrays in the past, I tend to use structured arrays instead of recarrays. The syntactic sugar isn't worth the trouble.
unutbu
2009-11-10 17:03:33
+4
A:
Use numpy.hstack():
>>> import numpy
>>> desc = {'names': ('gender','age','weight'), 'formats': ('S1', 'f4', 'f4')}
>>> a = numpy.array([('M',64.0,75.0),('F',25.0,60.0)], dtype=desc)
>>> numpy.hstack((a,a))
array([('M', 64.0, 75.0), ('F', 25.0, 60.0), ('M', 64.0, 75.0),
('F', 25.0, 60.0)],
dtype=[('gender', '|S1'), ('age', '<f4'), ('weight', '<f4')])
rcs
2009-11-10 15:39:36