comparing two arrays

[Sheldon]

Hi,

I have two arrays that are identical and contain 1s and zeros. Only the
ones are valid and I need to know where both arrays have ones in the
same position. I thought logical_and would work but this example proves
otherwise:

a = [0,1,2,5,6,6]
b = [5,4,1,6,4,6]
Numeric.logical_and(a==6,b==6) 0
Numeric.where(a==b,1,0) 0
Numeric.where(a==6 and b==6,1,0)

0

The where() statement is also worhtless here. Does anyone have any
suggestion on how to do this?

Thanks in advance,
Sheldon

 

[Diez B. Roggisch]

Hi,

I have two arrays that are identical and contain 1s and zeros. Only the

Obviously they aren't identical. They may be of same size.

ones are valid and I need to know where both arrays have ones in the
same position. I thought logical_and would work but this example proves
otherwise:

a = [0,1,2,5,6,6]
b = [5,4,1,6,4,6]
Numeric.logical_and(a==6,b==6) 0
Numeric.where(a==b,1,0) 0
Numeric.where(a==6 and b==6,1,0)

0

The where() statement is also worhtless here. Does anyone have any
suggestion on how to do this?

print [i for i, _ in enumerate((None for v in zip(a, b) where v == (1,1)))]

should give you the list of indices.

Diez

 

[Bas]

You are comparing a normal python list to a constant, which are
obviously unequal. Try converting your lists to arrays first
(untested):

import numeric/numpy as N
a =N.array([0,1,2,5,6,6])
b = N.array([5,4,1,6,4,6])
print a==6 and b==6
print N.where(a==6 and b==6)

hth,
Bas

Hi,

I have two arrays that are identical and contain 1s and zeros. Only the
ones are valid and I need to know where both arrays have ones in the
same position. I thought logical_and would work but this example proves
otherwise:

a = [0,1,2,5,6,6]
b = [5,4,1,6,4,6]
Numeric.logical_and(a==6,b==6) 0
Numeric.where(a==b,1,0) 0
Numeric.where(a==6 and b==6,1,0)

0

The where() statement is also worhtless here. Does anyone have any
suggestion on how to do this?

Thanks in advance,
Sheldon

 

[Diez B. Roggisch]

print [i for i, _ in enumerate((None for v in zip(a, b) where v ==
(1,1)))]

should give you the list of indices.

I musunderstood your question. Use


print [i for i, _ in enumerate((None for x, y in zip(a, b) where x == y))]

instead.

Diez

 

[Robert Kern]

You are comparing a normal python list to a constant, which are
obviously unequal. Try converting your lists to arrays first
(untested):

import numeric/numpy as N
a =N.array([0,1,2,5,6,6])
b = N.array([5,4,1,6,4,6])
print a==6 and b==6
print N.where(a==6 and b==6)

Careful there. The "and" keyword cannot be overloaded and so neither Numeric nor
numpy does. Either N.logical_and() should be used or (since the results of a==6
and b==6 are known to be boolean arrays) the & operator works fine as well.


In [9]: import numpy as np

In [10]: a = np.array([0,1,2,5,6,6])

In [11]: b = np.array([5,4,1,6,4,6])

In [12]: (a==6) & (b==6)
Out[12]: array([False, False, False, False, False, True], dtype=bool)

In [13]: np.where((a==6) & (b==6))
Out[13]: (array([5]),)


The OP may also find that numpy questions are best handled on numpy-discussion
rather than comp.lang.python .

https://lists.sourceforge.net/lists/listinfo/numpy-discussion

--
Robert Kern

"I have come to believe that the whole world is an enigma, a harmless enigma
that is made terrible by our own mad attempt to interpret it as though it had
an underlying truth."
-- Umberto Eco

 

[Sheldon]

Diez B. Roggisch skrev:

print [i for i, _ in enumerate((None for v in zip(a, b) where v ==
(1,1)))]

should give you the list of indices.

I musunderstood your question. Use


print [i for i, _ in enumerate((None for x, y in zip(a, b) where x == y))]

instead.

Diez

Hi Diez,

I wish I say that I understood what you wrote here but I can't.
Do you mind explaining a little more?

/sheldon

 

[Diez B. Roggisch]

print [i for i, _ in enumerate((None for x, y in zip(a, b) where x ==

y))]

instead.

Diez

Hi Diez,

I wish I say that I understood what you wrote here but I can't.
Do you mind explaining a little more?

I actually made a typo, instead of "where" in the above use "if". and the
whole thing won't compute what you are after. This will:

[i for i, equals in enumerate((x == y for x, y in zip(a, b))) if equals]

Sorry for that.

Its a nested list-comprehension. Actually, the outer thingy is a
list-comprehension while the inner is a generator expression. There is of
course a difference, but for now this doesn't matter too much.

a list comprehension of form

[<expression> for <variable> in <iterable> if <condition>]

is a more compact form for

result = []
for <variable> in <iterable>:
if <condition>:
result.append(<expression>)


A generator expression works the same with () instead of [].

Now let's decompose the above statment:

(x == y for x, y in zip(a, b) ))

<variable> = x, y
<iterable> = zip(a, b)
<expression> = x == y

So: the iterable is returned by the function zip. That is a built-in which
will take 2 or more lists and return a list of tuples, where the first
element is from the first list, then the second and so on. So your example
lists become:

a = [0,1,2,5,6,6]
b = [5,4,1,6,4,6]
zip(a, b) = [(0,5), (1,4), (2,1), (5,6), (6,4), (6,6)]

So iterating over this yields the pairs of (0,5) and so forth.

Next thing is that with

x, y = p

python unpacks a tuple and refers to its contents by name x for the first
one and y for the second. So overall, we loop in sync over both lists, and
getting x and y to point to the corresponding elements.

Now the expression x == y will result True if x == y - False otherwise.

So the result of the inner listcomp/genexp looks like this:

res = [False, False, False, False, False, True]

As you can see: for each pair x,y in the original lists we come up with the
result of comparing them.

Now the outer listcomp using "res" instead of the genexps for clarity reads
like this:



[i for i, equals in enumerate(res) if equals]

<variable> = i, equals
<iterable> = enumerate(res)
<expression> = i
<condition> = equals

enumerate is another built-in that takes an iterable and returns a tuple of

(pos, element)

for each element in the iterable.

So for our list, it will return:

[(0, False), (1, False), (2, False), (3, False), (4, False), (5, True)]

These tupkes values are assigened to i and equals for each element of the
above list. The condtion

equals

will then guarantee that only those expressions are evaluated where equals
is True - the last pair, so to speak. The expression then only stores the
index i. Which will give us the desired result.

Diez

 

[Maric Michaud]

Le Mardi 20 Juin 2006 12:09, Diez B. Roggisch a écrit :

[i for i, equals in enumerate((x == y for x, y in zip(a, b))) if equals]

No needs to nest comprehensions, should be :

[ i for i, v in enumerate(zip(a, b)) if v[0] == v[1] ]


--
_____________

Maric Michaud
_____________

Aristote - www.aristote.info
3 place des tapis
69004 Lyon
Tel: +33 426 880 097

 

[Diez B. Roggisch]

Le Mardi 20 Juin 2006 12:09, Diez B. Roggisch a écrit :

[i for i, equals in enumerate((x == y for x, y in zip(a, b))) if equals]

No needs to nest comprehensions, should be :

[ i for i, v in enumerate(zip(a, b)) if v[0] == v[1] ]

You're right, that design stemmed from my first broken version.

Diez

 

[Scott David Daniels]

Le Mardi 20 Juin 2006 12:09, Diez B. Roggisch a écrit :

[i for i, equals in enumerate((x == y for x, y in zip(a, b))) if equals]

No needs to nest comprehensions, should be :
[ i for i, v in enumerate(zip(a, b)) if v[0] == v[1] ]

You're right, that design stemmed from my first broken version.

Or even deconstruct to avoid the (very mildly confusing) v[0], v[1]:

[i for i, (left, right) in enumerate(zip(a, b)) if left == right]

 

[Sheldon]

Thanks Diez,

It will take a little while for this one to sink in but it gets the job
done now and will for future cases.

/Sheldon

Diez B. Roggisch skrev:

print [i for i, _ in enumerate((None for x, y in zip(a, b) where x ==
y))]

instead.

Diez

Hi Diez,

I wish I say that I understood what you wrote here but I can't.
Do you mind explaining a little more?

I actually made a typo, instead of "where" in the above use "if". and the
whole thing won't compute what you are after. This will:

[i for i, equals in enumerate((x == y for x, y in zip(a, b))) if equals]

Sorry for that.

Its a nested list-comprehension. Actually, the outer thingy is a
list-comprehension while the inner is a generator expression. There is of
course a difference, but for now this doesn't matter too much.

a list comprehension of form

[<expression> for <variable> in <iterable> if <condition>]

is a more compact form for

result = []
for <variable> in <iterable>:
if <condition>:
result.append(<expression>)


A generator expression works the same with () instead of [].

Now let's decompose the above statment:

(x == y for x, y in zip(a, b) ))

<variable> = x, y
<iterable> = zip(a, b)
<expression> = x == y

So: the iterable is returned by the function zip. That is a built-in which
will take 2 or more lists and return a list of tuples, where the first
element is from the first list, then the second and so on. So your example
lists become:

a = [0,1,2,5,6,6]
b = [5,4,1,6,4,6]
zip(a, b) = [(0,5), (1,4), (2,1), (5,6), (6,4), (6,6)]

So iterating over this yields the pairs of (0,5) and so forth.

Next thing is that with

x, y = p

python unpacks a tuple and refers to its contents by name x for the first
one and y for the second. So overall, we loop in sync over both lists, and
getting x and y to point to the corresponding elements.

Now the expression x == y will result True if x == y - False otherwise.

So the result of the inner listcomp/genexp looks like this:

res = [False, False, False, False, False, True]

As you can see: for each pair x,y in the original lists we come up with the
result of comparing them.

Now the outer listcomp using "res" instead of the genexps for clarity reads
like this:



[i for i, equals in enumerate(res) if equals]

<variable> = i, equals
<iterable> = enumerate(res)
<expression> = i
<condition> = equals

enumerate is another built-in that takes an iterable and returns a tuple of

(pos, element)

for each element in the iterable.

So for our list, it will return:

[(0, False), (1, False), (2, False), (3, False), (4, False), (5, True)]

These tupkes values are assigened to i and equals for each element of the
above list. The condtion

equals

will then guarantee that only those expressions are evaluated where equals
is True - the last pair, so to speak. The expression then only stores the
index i. Which will give us the desired result.

Diez

 

[Sheldon]

Thanks Diez,

It will take a little while for this one to sink in but it gets the job
done now and will for future cases.

/Sheldon

Diez B. Roggisch skrev:

print [i for i, _ in enumerate((None for x, y in zip(a, b) where x ==
y))]

instead.

Diez

Hi Diez,

I wish I say that I understood what you wrote here but I can't.
Do you mind explaining a little more?

I actually made a typo, instead of "where" in the above use "if". and the
whole thing won't compute what you are after. This will:

[i for i, equals in enumerate((x == y for x, y in zip(a, b))) if equals]

Sorry for that.

Its a nested list-comprehension. Actually, the outer thingy is a
list-comprehension while the inner is a generator expression. There is of
course a difference, but for now this doesn't matter too much.

a list comprehension of form

[<expression> for <variable> in <iterable> if <condition>]

is a more compact form for

result = []
for <variable> in <iterable>:
if <condition>:
result.append(<expression>)


A generator expression works the same with () instead of [].

Now let's decompose the above statment:

(x == y for x, y in zip(a, b) ))

<variable> = x, y
<iterable> = zip(a, b)
<expression> = x == y

So: the iterable is returned by the function zip. That is a built-in which
will take 2 or more lists and return a list of tuples, where the first
element is from the first list, then the second and so on. So your example
lists become:

a = [0,1,2,5,6,6]
b = [5,4,1,6,4,6]
zip(a, b) = [(0,5), (1,4), (2,1), (5,6), (6,4), (6,6)]

So iterating over this yields the pairs of (0,5) and so forth.

Next thing is that with

x, y = p

python unpacks a tuple and refers to its contents by name x for the first
one and y for the second. So overall, we loop in sync over both lists, and
getting x and y to point to the corresponding elements.

Now the expression x == y will result True if x == y - False otherwise.

So the result of the inner listcomp/genexp looks like this:

res = [False, False, False, False, False, True]

As you can see: for each pair x,y in the original lists we come up with the
result of comparing them.

Now the outer listcomp using "res" instead of the genexps for clarity reads
like this:



[i for i, equals in enumerate(res) if equals]

<variable> = i, equals
<iterable> = enumerate(res)
<expression> = i
<condition> = equals

enumerate is another built-in that takes an iterable and returns a tuple of

(pos, element)

for each element in the iterable.

So for our list, it will return:

[(0, False), (1, False), (2, False), (3, False), (4, False), (5, True)]

These tupkes values are assigened to i and equals for each element of the
above list. The condtion

equals

will then guarantee that only those expressions are evaluated where equals
is True - the last pair, so to speak. The expression then only stores the
index i. Which will give us the desired result.

Diez