使用perl 统计相关问题

xiaoao7678

[2012-07-11 15:49:28] thread:2144002 src:a
[2012-07-11 15:49:28] thread:2144002 src:b
[2012-07-11 15:49:28] thread:2144002 src:c
[2012-07-11 15:49:28] thread:2144002 src:a
[2012-07-11 16:42:02] thread:2144002 src:a

我想统计每小时，src 后面的数字有多少。求助各路高人
最终实现如
time   a   b   c
15      2   1   1
16     1    0   0

kk861123

1. #!/usr/bin/perl -w
   
2. use strict;
   
3. my %hour_of;
   
4. my %src;
   
5. while (<DATA>) {
   
6.     if (m{^\[\S+\s+(\d+):.*\]\s+\S+\s+src:(.)$}) {
   
7.         $hour_of{$1}{$2}++;
   
8.         $src{$2}++;
   
9.     }
   
10. }
    
11. 
    
12. print join("\t",'time'
    
13.                ,sort keys %src )
    
14.                ,"\n";
    
15. for my $h (sort keys %hour_of) {
    
16.     print join("\t",$h
    
17.                    ,map { $hour_of{$h}->{$_} || 0 } sort keys %src )
    
18.                    ,"\n";
    
19. }
    
20. 
    
21. __DATA__
    
22. [2012-07-11 15:49:28] thread:2144002 src:a
    
23. [2012-07-11 15:49:28] thread:2144002 src:b
    
24. [2012-07-11 15:49:28] thread:2144002 src:c
    
25. [2012-07-11 15:49:28] thread:2144002 src:a
    
26. [2012-07-11 16:42:02] thread:2144002 src:a
    

复制代码

cdtits

1. #!/usr/bin/env perl
   
2. 
   
3. use List::MoreUtils qw(uniq);
   
4. 
   
5. do { if (/^\[\S+ (\d+):\d+:\d+\] \w+:\d+ src:(\w+)$/) { $tmp{$1}->{$2}++; } } for (<DATA>);
   
6. @keys = map { keys %{$_} } values %tmp;
   
7. @keys = sort(uniq(@keys));
   
8. print "time\t", join("\t", @keys), "\n";
   
9. for $k (sort keys %tmp) {
   
10.     print "$k\t", join("\t", map { ${$tmp{$k}}{$_} || 0; } @keys), "\n";
    
11. }
    
12. 
    
13. __DATA__
    
14. [2012-07-11 15:49:28] thread:2144002 src:a
    
15. [2012-07-11 15:49:28] thread:2144002 src:b
    
16. [2012-07-11 15:49:28] thread:2144002 src:c
    
17. [2012-07-11 15:49:28] thread:2144002 src:a
    
18. [2012-07-11 16:42:02] thread:2144002 src:a

复制代码

xiaoao7678

回复 2# kk861123

你好，我的最终实现如下。15点 a 出现了几次，b出现了几次，c出现了几次
time   a   b   c
15      2   1   1
16     1    0   0

xiyoulaoyuanjia

1. #!/usr/bin/perl -w
   
2. 
   
3. use warnings;
   
4. my %time;
   
5. while(<DATA>){
   
6.         if(/\[\d*\-\d*\-\d*\s+(\d*)\:\d*\:\d*\]\s\w*\:\d*\s\w*\:(\w)/){
   
7.                 $time{$1}->{$2}++;
   
8.         }
   
9. }
   
10. 
    
11. print "Time\ta\tb\tc\n";
    
12. print $_."\t".($time{$_}{'a'}||0)."\t".($time{$_}{'b'}||0)."\t".($time{$_}{'c'}||0)."\n" for sort(keys %time);
    
13. 
    
14. __DATA__
    
15. [2012-07-11 15:49:28] thread:2144002 src:a
    
16. [2012-07-11 15:49:28] thread:2144002 src:b
    
17. [2012-07-11 15:49:28] thread:2144002 src:c
    
18. [2012-07-11 15:49:28] thread:2144002 src:a
    
19. [2012-07-11 16:42:02] thread:2144002 src:a
    

复制代码

xiaoao7678

谢谢你们，用你们的方法都得到了解决。

kk861123

回复 4# xiaoao7678


    sorry，少了一个sort在line17，已修改。

rubyish

嗯。。嗯。。
非常感谢！各路高人谢谢你们，

rubyish

我的最终实现如下:

1. #!perl
   
2. 
   
3. use 5.016;
   
4. my ( %h, %k );
   
5. /\s(\d+).*(\w+)(?{$h{$1}{$2}++})/ while <DATA>;
   
6. my @k = sort grep { !$k{$_}++ } map { keys $_ } values %h;
   
7. say join "\t", 'time', @k;
   
8. for my $t ( sort keys %h ) {
   
9.     say join "\t", $t, map { $h{$t}{$_} // 0 } @k;
   
10. }

复制代码

zhlong8

我以为正常人会把时间表示成对象，然后想统计什么统计什么