slowdown after each loop

**bvdet** · Feb 25 '08, 10:32 PM

Originally posted by newmaid

Hi all

just put together this little bit of code... I am sure it is heavily flawed:(... but can someone please point out to me the reason why it becomes slower after every loop. thanks!!!

** ....i am using "globals()[]" to try and create dynamic variables.. i am guessing that this a very bad way to do it?

Code:

j = 0

while j < x:

    s = 0
    
    while s < len(securityNamesList):
        
        globals()[securityNamesList[s] + 'adjustedHPR'] = []

        k = 0
        while k <  len(globals()[securityNamesList[0] + 'dollarHPRS']):
          
                              
            globals()[securityNamesList[s] + 'adjustedHPR'].append(globals()[securityNamesList[s] + 'FinalDistribution'][j] * globals()[securityNamesList[s] + 'dollarHPRS'][k])
            k = k + 1
       
        s = s + 1

    k = 0
    while k < len(globals()[securityNamesList[0] + 'adjustedHPR']):    
        s = 0
        todaysHPR = 0
        while s < len(securityNamesList):
        
            todaysHPR = todaysHPR + globals()[securityNamesList[s] + 'adjustedHPR'][k] 
            
            s = s + 1

        portfolioHPR.append( 1 + todaysHPR )
        k = k + 1


    portfolioDailyMeanReturn = stats.lmean(portfolioHPR)
    portfolioDailyStandardDeviation = stats.stdev(portfolioHPR)
    portfolioAnnualStandardDeviation = portfolioDailyStandardDeviation * 16
    portfolioDailyEstimatedGeometricMean = math.sqrt(portfolioDailyMeanReturn ** 2 - portfolioDailyStandardDeviation ** 2 )
    portfolioAnnualEstimatedGemoetricMean = portfolioDailyEstimatedGeometricMean ** 256
    portfolioSharpeRatio = (portfolioAnnualEstimatedGemoetricMean - 1) / portfolioAnnualStandardDeviation
    sharpeRatioList.append(portfolioSharpeRatio)
    

    

    j = j + 1

thanks again!

It's hard for me to follow without seeing some of the data objects you are processing. Try creating a separate dictionary dd instead of manipulating the global dictionary. When you are finished, you can update the global dictionary: globals().updat e(dd)

I prefer to use the augmented assignment operator.[code=Python]k +=1
# instead of
k = k+1[/code]It won't affect the speed though. You may be able to calculate todaysHPR like this:[code=Python]todaysHPR = sum([globals()[name + 'adjustedHPR'][k] for name in securityNamesLi st])[/code]

**newmaid** · Feb 26 '08, 09:09 AM

Originally posted by bvdet

It's hard for me to follow without seeing some of the data objects you are processing. Try creating a separate dictionary dd instead of manipulating the global dictionary. When you are finished, you can update the global dictionary: globals().updat e(dd)

I prefer to use the augmented assignment operator.[code=Python]k +=1
# instead of
k = k+1[/code]It won't affect the speed though. You may be able to calculate todaysHPR like this:[code=Python]todaysHPR = sum([globals()[name + 'adjustedHPR'][k] for name in securityNamesLi st])[/code]

thanks for your response bvdet. I think i found the problem. the portfolioHPR list was not being emptied after each loop .. and so was just getting bigger and bigger.

i have added [code=python] del porftolioHPR[:] [/code] after the calcs and it seems to have stopped the problem.

in the meantime I will try all your tips

thanks again

slowdown after each loop

slowdown after each loop

Comment

Comment