Laying multi-level reparations without losing their repeatability

Question

Suppose I do two repetitions with the same type and add them:

>>> import numpy as np >>> dt = [('foo', int), ('bar', float)] >>> a = np.empty(2, dtype=dt).view(np.recarray) >>> b = np.empty(3, dtype=dt).view(np.recarray) >>> c = np.hstack((a,b)) 

Although a and b are recursions, c not:

 >>> c.foo Traceback (most recent call last): File "<stdin>", line 1, in <module> AttributeError: 'numpy.ndarray' object has no attribute 'foo' >>> d = c.view(np.recarray) >>> d.foo array([ 0, 111050731618561, 0, 7718048, 8246760947200437872]) 

I can obviously include it again in the recheck, as shown above with d , but this is inconvenient. Is there a reason why stacking two recursive data does not lead to another repetition?

Answer 1

I dont know. This is most likely an error / function that has never been implemented. numpy.hstack is basically a wrapper around a function in numpy.core.fromnumeric . Numeric is one of two numpy predecessors. Most functions in numpy have an agreement for output of the same type as input, calling the __array_wrap__ input method on the output, and the resulting output should have the same data, but "wrapped" in a new class. Perhaps the concept of “wrapping” was not in numerical form and was never added to this function.

You can use this technique to make smarter styling function.

 def hstack2(arrays) : return arrays[0].__array_wrap__(numpy.hstack(arrays)) 

This works for both recursive and regular arrays.

 >>> f = hstack2((a,b)) >>> type(f) <class 'numpy.core.records.recarray'> >>> f.foo array([ 140633760262784, 111050731618561, 140633760262800, 7536928, 8391166428122670177]) >>> x = numpy.random.rand(3) >>> y = numpy.random.rand(2) >>> z = hstack2((x,y)) >>> type(z) <type 'numpy.ndarray'> 

I'm not sure what you are planning, but you can ask the numpy mailing list there is a better way than using the documented but bidirectional method and that their arguments are that they do not wrap.

Answer 2

As an alternative, there are helper utilities numpy.lib.recfunctions that I came across. This module has functions for merging and stacking recarrays :

 from numpy.lib.recfunctions import stack_arrays c = stack_arrays((a, b), asrecarray=True, usemask=False) c.foo >>> array([ 140239282560000, 4376479720, -4611686018427387904, 4358733828, 4365061216]) 

If you want to add additional columns to recarray , this can be done with merge_arrays :

 import numpy as np from numpy.lib.recfunctions import merge_arrays dt1 = [('foo', int), ('bar', float)] dt2 = [('foobar', int), ('barfoo', float)] aa = np.empty(6, dtype=dt1).view(np.recarray) bb = np.empty(6, dtype=dt2).view(np.recarray) cc = merge_arrays((aa, bb), asrecarray=True, flatten=True) type(cc) >>> numpy.core.records.recarray 

(Although this is not an answer to the question, I am posting the last example as a reference)

Answer 3

By the way, you can also use:

 c = np.concatenate((a,b)) 

or

 c = np.r_[a, b] 

(Source: This is a mailing list message )