Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: MATA function modifies global variables


From   George Vega Yon <[email protected]>
To   [email protected]
Subject   Re: st: MATA function modifies global variables
Date   Fri, 7 Jun 2013 09:30:02 -0400

Now I get why the performance other functions didn't change after I
replace several variables by pointers... Great news though! A must
know for MATA users.

Thanks!
George Vega Yon
7 647 2552
http://cl.linkedin.com/in/georgevegayon


2013/6/6 Sergiy Radyakin <[email protected]>:
> That's right. See the chapter "The by-address calling convention" in
> the mata manual and a quote:
>
> "Arguments are passed to functions by address, not by value. If you
> change the value of an argument,
> you will change the caller’s argument."
>
> Here are some simpler examples:
> mata mata clear
> mata
>     Y=1,2,3,4,5
>     real matrix foobar(real matrix y) {
>       y=y'
>       return(y)
>     }
>
>     Y
>     foobar(Y)
>     foobar(Y)
>
>
>     S="abcdefg"
>     strlen(S)
>     string foobars(string s) {
>       s=substr(s,1,strlen(s)-1)
>       return(s)
>     }
>
>     S
>     foobars(S)
>     foobars(S)
>
>
>     N=17
>     real foobarn(real n) {
>       n=n-1
>       return(n)
>     }
>
>     foobarn(N)
>     foobarn(N)
>
>
> end
>
>
>
> Best, Sergiy
>
> On Thu, Jun 6, 2013 at 4:23 PM, George Vega Yon <[email protected]> wrote:
>> Sorry, I forgot to change that, blop_vec() is "myfun" and neig is "X".
>> I'll just rewrite the code with the corrections. Thank you!
>>
>> __________________________
>> mata:
>> /* Function definition */
>> real colvector myfun(real colvector y, real matrix x) {
>>
>> real matrix z
>>
>> y = y'
>> x = x'
>>
>> ... some code lines (no poiter to Y or X)...
>>
>> return(z)
>> }
>>
>> X = (
>>         0,0 \
>>         1,0 \
>>         1,1 \
>>         0,1 \
>>         1/2,1/2\
>>         .9,1/2 \
>>         .2,.5 \
>>         .8,.3
>>         )
>> Y = runiform(rows(X),1)
>>
>>
>> /* Testing */
>> timer_clear()
>> for(i=1;i<=100;i++) {
>>         timer_on(1)
>>         YC = myfun(Y,X)
>>         timer_off(1)
>> }
>> timer()
>>
>> end
>> __________________________
>> George Vega Yon
>> 7 647 2552
>> http://cl.linkedin.com/in/georgevegayon
>>
>>
>> 2013/6/6 Sergiy Radyakin <[email protected]>:
>>> Can you tell us what neig is? what blop_vec is?
>>>
>>> On Thu, Jun 6, 2013 at 4:10 PM, George Vega Yon <[email protected]> wrote:
>>>> Dear Statalisters,
>>>>
>>>> I'm having a weird problem on variables' scope. I have a function that
>>>> takes two objects (a real colvector and a matrix) wich, once inside
>>>> the function, are transposed. This function returns a matrix (builded
>>>> inside of the function of course). The problem is that the function
>>>> seems to actually transpose the global objects (not only the copies
>>>> within the function), thing that, as far as I know, should not be
>>>> happening as I'm not using pointers to the global objects.
>>>>
>>>> My code looks something like this
>>>> __________________________
>>>> mata:
>>>> /* Function definition */
>>>> real colvector myfun(real colvector y, real matrix x) {
>>>>
>>>> real matrix z
>>>>
>>>> y = y'
>>>> x = x'
>>>>
>>>> ... some code lines (no poiter to Y or X)...
>>>>
>>>> return(z)
>>>> }
>>>>
>>>> Y = runiform(rows(neig),1)
>>>> X = (
>>>>         0,0 \
>>>>         1,0 \
>>>>         1,1 \
>>>>         0,1 \
>>>>         1/2,1/2\
>>>>         .9,1/2 \
>>>>         .2,.5 \
>>>>         .8,.3
>>>>         )
>>>>
>>>>
>>>> /* Testing */
>>>> timer_clear()
>>>> for(i=1;i<=100;i++) {
>>>>         timer_on(1)
>>>>         YC = blop_vec(Y,X)
>>>>         timer_off(1)
>>>> }
>>>> timer()
>>>>
>>>> end
>>>> __________________________
>>>>
>>>>
>>>> I get the error
>>>>
>>>> __________________________
>>>> myfun:  3203  Y[1,8] found where colvector required
>>>>                  <istmt>:     -  function returned error
>>>> __________________________
>>>>
>>>> Thanks in advanced,
>>>>
>>>> George Vega Yon
>>>> 7 647 2552
>>>> http://cl.linkedin.com/in/georgevegayon
>>>> *
>>>> *   For searches and help try:
>>>> *   http://www.stata.com/help.cgi?search
>>>> *   http://www.stata.com/support/faqs/resources/statalist-faq/
>>>> *   http://www.ats.ucla.edu/stat/stata/
>>> *
>>> *   For searches and help try:
>>> *   http://www.stata.com/help.cgi?search
>>> *   http://www.stata.com/support/faqs/resources/statalist-faq/
>>> *   http://www.ats.ucla.edu/stat/stata/
>> *
>> *   For searches and help try:
>> *   http://www.stata.com/help.cgi?search
>> *   http://www.stata.com/support/faqs/resources/statalist-faq/
>> *   http://www.ats.ucla.edu/stat/stata/
>
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/faqs/resources/statalist-faq/
> *   http://www.ats.ucla.edu/stat/stata/

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/


© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index