[Numpy-discussion] [SciPy-user] unique, sort, sortrows

Robert Cimrman cimrman3@ntc.zcu...
Sat Jul 26 16:35:24 CDT 2008

Hi David,

I can comment on unique1d, as I am the culprit. I am cc'ing to  
numpy-discussion as this is a numpy function.

Quoting "David M. Kaplan" <David.Kaplan@ird.fr>:
> 2) Is there a simple equivalent of sortrows(a) (i.e., sorting by entire
> rows)?  Similarly, is there a simple equivalent of the matlab Y =

have you looked at lexsort?

> 3) Is there an equivalent of [Y,I,J] = unique(X)?  In this case, I is
> the indices of the unique elements and J is an index from Y to X (i.e.,
> where the unique elements appear in X.  I can get I with:
> I,Y = unique1d( X, return_index=True )
> but J, which is very useful, is not available.  I suppose I could do:
> J = array([])
> for i,y in enumerate(Y):
> 	J[ nonzero( y == X ) ] = i
> But it seems like J is useful enough that there should be an easier way
> (or perhaps it should be integrated into unique1d, at the risk of adding
> more keyword arguments).

So basically Y = X[I] and X = Y[J], right? I do not recall matlab that  
well to know for sure. It certainly could be done, I could look at it  
after I return from Euroscipy (i.e. after Monday).

I would replace return_index argument by two arguments: return_direct  
(->I) and return_inverse (->J), ok? Does anyone propose better names?

Actually most of the methods in arraysetops could return optionally  
some index arrays. Would anyone use it? (I do not need it personally :)


More information about the Numpy-discussion mailing list