Velocity Reviews > C++ > combining two vectors

# combining two vectors

Alan
Guest
Posts: n/a

 11-14-2006
Is there an easy and efficient way to combine two <vector>s, rather
than taking each element from one and adding it to the other?

I haven`t been able to find any guidance on or examples of this
sort of operation.

Thanks, Alan

Ian Collins
Guest
Posts: n/a

 11-14-2006
Alan wrote:
> Is there an easy and efficient way to combine two <vector>s, rather
> than taking each element from one and adding it to the other?
>
> I haven`t been able to find any guidance on or examples of this
> sort of operation.
>

vector1.insert( vector1.end(), vector2.begin(), vector2.end() );

--
Ian Collins.

Alan
Guest
Posts: n/a

 11-14-2006
Ian, Thanks! Alan

Stefan Naewe
Guest
Posts: n/a

 11-15-2006
Ian Collins schrieb:
> Alan wrote:
>> Is there an easy and efficient way to combine two <vector>s, rather
>> than taking each element from one and adding it to the other?
>>
>> I haven`t been able to find any guidance on or examples of this
>> sort of operation.
>>

>
> vector1.insert( vector1.end(), vector2.begin(), vector2.end() );
>

It's easy - yes.
But is it really efficient ?
Isn't it as unefficient as repeatedly calling push_bask() on vector1 ?

/S
--
Stefan Naewe
stefan_DOT_naewe_AT_atlas_DOT_de

Pete Becker
Guest
Posts: n/a

 11-15-2006
Stefan Naewe wrote:
> Ian Collins schrieb:
>> Alan wrote:
>>> Is there an easy and efficient way to combine two <vector>s, rather
>>> than taking each element from one and adding it to the other?
>>>
>>> I haven`t been able to find any guidance on or examples of this
>>> sort of operation.
>>>

>> vector1.insert( vector1.end(), vector2.begin(), vector2.end() );
>>

>
> It's easy - yes.
> But is it really efficient ?
> Isn't it as unefficient as repeatedly calling push_bask() on vector1 ?
>

No. The iterators that represent the range being copied are random
access iterators, so insert can figure out how many elements will be
could end up reallocating the storage space more than once.

--

-- Pete
Roundhouse Consulting, Ltd. (www.versatilecoding.com)
Author of "The Standard C++ Library Extensions: a Tutorial and
Reference." (www.petebecker.com/tr1book)

Stefan Naewe
Guest
Posts: n/a

 11-15-2006
Pete Becker schrieb:
> Stefan Naewe wrote:
>> Ian Collins schrieb:
>>> Alan wrote:
>>>> Is there an easy and efficient way to combine two <vector>s, rather
>>>> than taking each element from one and adding it to the other?
>>>>
>>>> I haven`t been able to find any guidance on or examples of this
>>>> sort of operation.
>>>>
>>> vector1.insert( vector1.end(), vector2.begin(), vector2.end() );
>>>

>>
>> It's easy - yes.
>> But is it really efficient ?
>> Isn't it as unefficient as repeatedly calling push_bask() on vector1 ?
>>

>
> No. The iterators that represent the range being copied are random
> access iterators, so insert can figure out how many elements will be
> could end up reallocating the storage space more than once.

OK.
But what if I do a vector1.reserve(x) (with a valid x...) before ?
'vector2.end()-vector2.begin()' elements need to be copied after vector1.end(), right?

(Using std::list and std::list::splice() would be more efficient in this case, wouldn't it?)

/S
--
Stefan Naewe
stefan_DOT_naewe_AT_atlas_DOT_de

red floyd
Guest
Posts: n/a

 11-15-2006
Stefan Naewe wrote:
> Ian Collins schrieb:
>> Alan wrote:
>>> Is there an easy and efficient way to combine two <vector>s, rather
>>> than taking each element from one and adding it to the other?
>>>
>>> I haven`t been able to find any guidance on or examples of this
>>> sort of operation.
>>>

>> vector1.insert( vector1.end(), vector2.begin(), vector2.end() );
>>

>
> It's easy - yes.
> But is it really efficient ?
> Isn't it as unefficient as repeatedly calling push_bask() on vector1 ?

Why does everyone ask "is this more efficient" in a vacuum?

Have you benchmarked to determine that your vector operations are the
bottleneck? An "efficient" program that doesn't work right is not as
good as an "inefficient" program that does.

Hoare's Law (also attributed to Knuth): "Premature optimization is the
root of all evil".

Pete Becker
Guest
Posts: n/a

 11-15-2006
red floyd wrote:
> Stefan Naewe wrote:
>> Ian Collins schrieb:
>>> Alan wrote:
>>>> Is there an easy and efficient way to combine two <vector>s, rather
>>>> than taking each element from one and adding it to the other?
>>>>
>>>> I haven`t been able to find any guidance on or examples of this
>>>> sort of operation.
>>>>
>>> vector1.insert( vector1.end(), vector2.begin(), vector2.end() );
>>>

>>
>> It's easy - yes.
>> But is it really efficient ?
>> Isn't it as unefficient as repeatedly calling push_bask() on vector1 ?

>
> Why does everyone ask "is this more efficient" in a vacuum?
>
> Have you benchmarked to determine that your vector operations are the
> bottleneck? An "efficient" program that doesn't work right is not as
> good as an "inefficient" program that does.
>

And how is changing from push_back to insert going to make this code not
work right?

> Hoare's Law (also attributed to Knuth): "Premature optimization is the
> root of all evil".

Okay, let's all use bubble sort until we can prove that quicksort will
be a better choice.

Sometimes optimization without measurement is perfectly appropriate.
When there's a choice of two ways to do something and one is slower,
choose the other.

for (iter = vector2.begin(); iter != vector2.end(); ++iter)
vector1.push_back(*iter);

versus

vector1.insert(vector1.end(), vector2.begin(), vector2.end());

No contest. Use the latter.

--

-- Pete
Roundhouse Consulting, Ltd. (www.versatilecoding.com)
Author of "The Standard C++ Library Extensions: a Tutorial and
Reference." (www.petebecker.com/tr1book)

Ian Collins
Guest
Posts: n/a

 11-15-2006
Pete Becker wrote:
> red floyd wrote:
>
>>
>> Why does everyone ask "is this more efficient" in a vacuum?
>>
>> Have you benchmarked to determine that your vector operations are the
>> bottleneck? An "efficient" program that doesn't work right is not as
>> good as an "inefficient" program that does.
>>

>
> And how is changing from push_back to insert going to make this code not
> work right?
>
>> Hoare's Law (also attributed to Knuth): "Premature optimization is the
>> root of all evil".

>
>
> Okay, let's all use bubble sort until we can prove that quicksort will
> be a better choice.
>
> Sometimes optimization without measurement is perfectly appropriate.
> When there's a choice of two ways to do something and one is slower,
> choose the other.
>

Agreed, maybe the quote should be changed to "Premature
micro-optimisation is the root of all evil" to differentiate between
code tweaks and choice of algorithm.

--
Ian Collins.

Alf P. Steinbach
Guest
Posts: n/a

 11-15-2006
* Ian Collins:
> Pete Becker wrote:
>> red floyd wrote:
>>
>>> Hoare's Law (also attributed to Knuth): "Premature optimization is the
>>> root of all evil".

>>
>> Okay, let's all use bubble sort until we can prove that quicksort will
>> be a better choice.
>>
>> Sometimes optimization without measurement is perfectly appropriate.
>> When there's a choice of two ways to do something and one is slower,
>> choose the other.
>>

> Agreed, maybe the quote should be changed to "Premature
> micro-optimisation is the root of all evil" to differentiate between
> code tweaks and choice of algorithm.

Isn't that an optimization?

--
A: Because it messes up the order in which people normally read text.
Q: Why is it such a bad thing?
A: Top-posting.
Q: What is the most annoying thing on usenet and in e-mail?