Yet another Python vs. Perl speed issue/question

[Christian Tanzer]

nanotech at europa.com wrote:

> Python 2.0 on an HP (nice and big) takes ~30 seconds to lowercase all
> the lines in a 3meg file. Perl takes less than a second. What am I
> doing wrong?!?
>
> Python
> ------
> import sys
> import string
>
> stdin=sys.stdin
> lower=string.lower
>
> print map(lower,stdin.readlines())
>
> Perl
> ----
>
> print map(lc,<STDIN>);

Your Python code is doing unnecessary work (BTW, it doesn't do what
you expect). Just for comparison, the timings for a 3MB file on my
aging Pentium Pro 200 MHz machine:

    $ time python /tmp/test1.py < /tmp/test.txt > /tmp/test.out 
    real    0m10.246s
    user    0m4.810s
    sys     0m1.190s
    $ time python /tmp/test2.py < /tmp/test.txt > /tmp/test.out 
    real    0m0.781s
    user    0m0.200s
    sys     0m0.270s

    #### test1.py                       #### test2.py
    import sys                          import sys
    import string                       import string
    stdin=sys.stdin                     stdin=sys.stdin
    lower=string.lower                  lower=string.lower
    print map(lower,stdin.readlines())  sys.stdout.write(lower(stdin.read()))

Your perl code gives a compilation error:

    $ time perl /tmp/test1.pl < /tmp/test.txt > /tmp/test.out
    Not enough arguments for lower case at /tmp/test1.pl line 1, near "lc,"
    Execution of /tmp/test1.pl aborted due to compilation errors.

Slightly changed, it lies halfway between the two python versions:

    $ time perl /tmp/test2.pl < /tmp/test.txt > /tmp/test.out
    real    0m3.297s
    user    0m2.830s
    sys     0m0.360s

    #### test2.pl
    print map {lc ($_)} <STDIN>;

-- 
Christian Tanzer                                         tanzer at swing.co.at
Glasauergasse 32                                       Tel: +43 1 876 62 36
A-1130 Vienna, Austria                                 Fax: +43 1 877 66 92