Question

我已经对这样的数据进行了排序（使用 ssgrep 从MS Excel转储到命令行）：

2012-06-27T22:55:00
2012-06-27T23:00:00
2012-06-27T23:05:00
2012-06-27T23:10:00
2012-06-27T23:15:00
2012-06-27T23:20:00
2012-06-27T23:25:00
2012-06-27T23:30:00

现在我想把它管道输入一个脚本（或任何命令行工具），它测试所有值是否确实是有效日期，以及它们之间是否有任何间隔不是5分钟。（我需要对每隔1小时采样的不同数据集做同样的事情）

是否有可用于命令行的工具可以执行此操作？我知道在R或Perl中编写一个小脚本很容易（事实上我已经开始编程了）。但也许我已经在linux / solaris上预装了一个我不知道的小工具。

Answer 1

我认为没有任何工具可以解决您的任务，但可以使用Perl / Python oneliner（或非常小的脚本）。

$ cat 1.pl
use Date::Parse;

while(<>) {
    next if /^\s*$/;
    $time = str2time($_);
    if ($time) {
        print $time,"\n";
        if ($last and $time - $last != 300) {
            print "delta is wrong: ".($last-$time)."\n";
        }
        $last = $time;
    }
    else {
        print "not a date","\n";
    }
}

$ cat 1.txt 
2012-06-27T22:55:00
2012-06-27T23:00:00
wrong
2012-06-27T23:05:00
2012-06-27T23:10:00
2012-06-27T23:15:01
2012-06-27T23:20:00
2012-06-27T23:25:00
2012-06-27T23:30:00

$ cat 1.txt | perl 1.pl
1340830500
1340830800
not a date
1340831100
1340831400
1340831701
delta is wrong: -301
1340832000
delta is wrong: -299
1340832300
1340832600

Answer 2

如果您已经使用过Bash，那么逻辑上您需要在脚本中继续使用Bash（如果可能的话）。

(your grep etc commands) | while read date; do
   curDate=$(date +%s -d "$date" 2> /dev/null);
   if [ "$curDate" != "" ]; then
      if [[ "$last" != "" && $((curDate - last)) -ne 300 ]]; then
         echo "wrong delta: " $((curDate - last));
      fi
      last=$curDate;
   else
      echo "not a date";
   fi
done

Answer 3

编辑：第二次尝试：

from datetime import datetime, timedelta
import dateutil.parser

# fake console input
input = "2012-06-27T22:55:00 \
2012-06-27T23:00:00 \
2012-06-27T23:05:00 \
2012-06-27T23:10:00 \
2012-06-27T23:15:00 \
2012-06-27T23:20:00 \
2012-06-27T23:25:00 \
2012-06-27T23:30:00"

def parse(s):
    """Returns datetime if parseable, otherwise None"""
    try:
        return dateutil.parser.parse(s)
    except ValueError:
        return None # not valid date string

# parse the times, remove Nones
dates = filter(None, (parse(i) for i in input.split()))

# pair every second datetime together
dates = zip(dates[0::2], dates[1::2])

# print every datetime pair that has a delta of exactly five minutes
for d in dates:
    if (d[1] - d[0]).seconds == (5*60):
        print d[0]
        print d[1]

Answer 4

这是我自己的重量级问题解决方案（正在进行的工作）。感谢所有贡献者。

#!/usr/bin/env perl
# This script checks if all input date values have a delta of 
# say, 5 minutes to the previous date value.
#
# $Id: test_datetintervals.t 1804 2012-08-07 10:35:48Z knb $
#
#  for f in $(ls -1 *std--ytd--2012-m01*xlsx); do echo ""; echo "$f"; ssgrep 2012- $f | test_datetintervals.t -num 60 -comm "$(basename $f)"; done
#

# ssgrep 2012 Monitoring_DailyAvg.min--2008-01-01--2012-11-01.txt.xlsx | grep -v "2012.15" | ./test_datetintervals.t -num 0 -interval day -pat \'%m/%d/%Y

use Modern::Perl;

use Getopt::Long;
use File::Basename;
use Carp qw (carp croak confess cluck);
use lib '/config';
use lib 'T:';
use lib '/lib/';    # wb3


use Pod::Usage;
use Test::More;
use DateTime;
use DateTime::Format::Strptime;

my $interval = "minute";                                                           # 2012-06-27T23:05:00
my $n        = 5;
my $comment = "";
my $help;
my $man;
my $pat = '%FT%T', #  2012-10-09T17:00:00
#my $pat = "'%m/%d/%Y", #  09/27/2012
my $argcnt = scalar(@ARGV);
GetOptions( 'num=i' => \$n, 'interval:s' => \$interval, 'comment:s' => \$comment,    'help|?'          => \$help,
    man               => \$man, 'pattern:s' => \$pat ) or pod2usage(2);
my $fn = basename($0);

my $msg2 = "This script must be part of a unix pipe. Script checks if all input date values have a delta of $n ${interval}s to the previous date value.";

pod2usage(
    -message    => "$msg2\nCall $fn  --man  to see full documentation",
    -exitstatus => 1
  )
  if $help;

my $msg = <<'MSG';
# Sample calls:
#
#  for f in $(ls -1 *std--ytd--2012.txt.xlsx); do echo ""; echo "$f"; ssgrep 2012- $f | test_datetintervals.t -num 60 -comm "$(basename $f)"; done
#

MSG

pod2usage( -message => "$msg2\n\n$msg", -exitstatus => 0 ) if $man;
pod2usage( -message => "# \n!!!! Script was called with no arguments- but this cannot work.\n\n $msg2\n\n$msg", -exitstatus => 0 ) if $argcnt == 0;

my $Strp = new DateTime::Format::Strptime(
    pattern   => $pat,
    locale    => 'de_DE',
    time_zone => 'floating',
    on_error  => \&my_carp
);

note($msg2);
my $line;
my $prev;
my $okflag = 1;
my $i = 0;
while (<>) {
    $line = $_;
    chomp $line;
    my $dt = $Strp->parse_datetime($line);

    if ( !$dt ) {
        note "Line $.: cannot parse '$line' to datetime object: $!";
    }
    if ( !$prev ) {
        note "Line $.: previous value not found/not applicable!";
        note $dt->datetime();
    } else {
        my $duration = $dt->delta_ms($prev);
        my $diff     = $duration->in_units("${interval}s");
        if ( $diff != $n ) {
            fail( "$.: $n $interval interval between " . $prev->datetime() . " and " . $dt->datetime() );
            note "prev: " . $prev->datetime();
            note "    : " . $dt->datetime();
            note "diff: $diff ${interval}s";
            note "";
            $okflag = 0;
        }
        $i++;

    }

    $prev = $dt;
}
if ($okflag){
    ok(1, "no gaps of $n ${interval}s detected! $comment");
    note "";
}
note("done testing '$comment', about $i lines");
note "";
done_testing();

sub my_carp {

    carp " line $.: Something wrong with " . $_;

    note  $_[1];
    return 1; # 0 for continue

}


=pod

usage: $fn.pl [-?] [long options...]
    -? --usage --help  Prints this usage information.
    --man               
    --num=i    # 5
    --interval # minutes           
    --comment           
    --pattern  #'%FT%T', #  2012-10-09T17:00:00
    --overwrite (or nooverwrite)
=cut

用于在bash命令行上测试日期时间间隔的工具/脚本

4 个答案: