[SciPy-User] [SciPy-user] Matching up arrays?

[Keith Goodman]

On Thu, Nov 18, 2010 at 9:58 PM, mdekauwe <mdekauwe at gmail.com> wrote:
>
> Hi,
>
> So I have 2 arrays which hold some data and for each array I have an array
> with the associated date stamp. The dates vary between arrays, so what I
> would like to be able to do is just subset the data so I end up with two
> arrays if there is a matching time stamp between arrays (with the idea being
> I would do some comparative stats on these arrays). How I solved it seems a
> bit ugly and I wondered if anyone had a better idea?
>
> e.g.
>
> m_date = np.array(['1998-01-01 00:00:00', '1999-01-01 00:00:00', '2000-01-01
> 00:00:00',
>                           '2005-01-01 00:00:00'])
> o_date = np.array(['1998-01-01 00:00:00', '1999-01-01 00:00:00', '2000-01-01
> 00:00:00'])
>
> mm = np.array([ 3.5732, 4.5761, 4.0994, 3.9031])
> oo = np.array([ 5.84, 5.66, 5.83])
>
> x, y = [], []
> o = np.vstack((o_date, oo))
> m = np.vstack((m_date, mm))
> for i in xrange(o.shape[1]):
>    for j in xrange(m.shape[1]):
>        if m[0,j] == o[0,i]:
>            x.append(m[1,j])
>            y.append(o[1,i])

You could try using a labeled-array (http://pypi.python.org/pypi/la).

Create larrys:

>> mlar = la.larry(mm, [m_date.tolist()])
>> olar = la.larry(oo, [o_date.tolist()])

Alignment is automatic for binary operations (the default is an inner join):

>> mlar + olar
label_0
    1998-01-01 00:00:00
    1999-01-01 00:00:00
    2000-01-01 00:00:00
x
array([  9.4132,  10.2361,   9.9294])

Our you can prealign the data using various join methods:

>> m, o = la.align(mlar, olar, join='outer')
>> o
label_0
    1998-01-01 00:00:00
    1999-01-01 00:00:00
    2000-01-01 00:00:00
    2005-01-01 00:00:00
x
array([ 5.84,  5.66,  5.83,   NaN])

If all you want are the underlying arrays, then:

>> m.A
   array([ 3.5732,  4.5761,  4.0994,  3.9031])
>> o.A
   array([ 5.84,  5.66,  5.83,   NaN])