Optimization and accelerating the loops in Python

Question

How do you optimize this three snippets of code? Especially the third because there are many combinations between the list values and the time is dangerously longer with 1000 inputs.

Code 1:

e00=[]
for i in range(len(c1)):
    for j in range(len(d1[i])):
        if d1[i][j]%2==0:
           d = [c1[i],d1[i][j]]
           e00.append(d)

Code 2:

sciezki=[]
for i in range(len(out2)):     
    x1 = out2[i][-len(out2[i])]
    x2 =out2[i][-1]    
    z1 = nx.shortest_path(g, x1, x2)  
    if z1 == out2[i] and len(z1)==8:
          sciezki.append(z1)

Code 3:

out=[] 
for h in range(len(k)):
    if len(out)!=0:
       k2 = [out, k[h]]
       for q in range(len(k2[0])):
           for w in range(len(k2[1])):
               r = list(chain(k2[0][q],k2[1][w]))
               p = [n for n, _ in groupby(r)]
               if len(p)==h+2:
                  out.append(p)  
    else:
       for i in range(len(k[0])):
           for j in range(len(k[1])):
               r = list(chain(k[0][i],k[1][j]))
               p = [n for n, _ in groupby(r)]
               if len(p)==3:
                   out.append(p)

Usually it is better to iterate over the list directly, so for ci in c, instead of for i in range(len(c)): — willeM_ Van Onsem
– willeM_ Van Onsem, Commented Nov 22, 2017 at 8:36
Start with analyzing the algorithm and see if you can make it faster. After that you turn to code details. — klutt
– klutt, Commented Nov 22, 2017 at 8:37
@klutt this is always the worst, the man sits a few days over the algorithm, and then you have to think about the possible algorithm change :)) — Tomasz Przemski
– Tomasz Przemski, Commented Nov 22, 2017 at 8:41
Well, without addressing the underlying algorithm I doubt you'll gain much in performance. But there are certain obvious things, like trying to iterate directly over a list rather than using indexing, and oh, don't do r = list(chain(...)) just use r = chain(...). And another micro-optimization of course is to "in-line" the .append method call: out = []; append_out = out.append then use append_out(p) in place of out.append(p). But all of these will be very minor, your problem is algorithmic complexity. — juanpa.arrivillaga
– juanpa.arrivillaga, Commented Nov 22, 2017 at 9:20

Alexander · Accepted Answer · 2017-11-22 08:58:16Z

1

Code 1

Use a conditional list comprehension together with enumeration:

e00 = [[c_val, d_val] 
       for i, c_val in enumerate(c1) 
       for d_val d1[i] 
       if not d_val % 2]

Code 2

Only slight modifications required.

g = ...  # Undefined in sample code.
sciezki = []
for out2_val in out2:     
    x1 = out2_val[-len(out2_val)]
    x2 = out2_val[-1]    
    z1 = nx.shortest_path(g, x1, x2)  
    if z1 == out2_val and len(z1) == 8:
        sciezki.append(z1)
        x = '\n'.join(map(str,sciezki))  # Remove if possible.

edited Nov 22, 2017 at 8:58

answered Nov 22, 2017 at 8:49

Alexander

111k32 gold badges212 silver badges208 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

Tomasz Przemski Over a year ago

Właściwie to 'x' tutaj jest niepotrzebne, mój punkt widzenia. 'g' to g = nx.Graph () z Networkx. 2 kod polega na porównaniu pierwszego i ostatniego węzła i przejścia przez "shortest_path", czy dany wybór jest jednoznaczny.

Collectives™ on Stack Overflow

Optimization and accelerating the loops in Python

1 Answer 1

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related