When running a long-term program, how to define the next program to run?

1) Whenever you run a long-term program (let us call it "prog.exe"), please do not run it directly.  You can create a batch file named "RunProg.bat" with following content:
----------------------
prog.exe
call next.bat
----------------------

2) Then, run the program with "RunProg.bat"

3) Now, you can define the next program to run freely and write them in "next.bat"

Check Date String
use HTTP::Date;
my $daystr="20091026";

if( str2time($daystr) )
{
   print "Yes";
}else
{
   print "No";
}

Today's String
my ($sec,$min,$hour,$mday,$mon,$year,$wday);
($sec,$min,$hour,$mday,$mon,$year,$wday,undef,undef)=localtime(time()); $year+=1900;$mon++;
 
$mon = "0".$mon unless $mon >= 10;
$mday = "0".$mday unless $mday >= 10;
 
my $today = "$year$mon$mday";

Iterate Days
for( my $str="20091001"; $str < "20101010"; $str++)
{
   if( str2time($str) )
   {
    #valid date string
   }
}

用NameManager代表离散地选择的几个列

 

=Sheet1!$B$1:$B$3,Sheet1!$J$1:$J$3,Sheet1!$A$1:$A$3

 

 

 

选择另一个表格的最后一行

行号'Newest Line'!$B$12 =  COUNT(Sheet1!A:A)

第一个值=OFFSET(Sheet1!$A$1,'Newest Line'!$B$12,COLUMN()-1)

其它值通过自动填充，自动添加。

Add Local Administrators
net localgroup Administrators "group\name" /add

Check Running Process
tasklist /m ie*

Creating

SET ANSI_NULLS ON
GO
SET QUOTED_IDENTIFIER ON
GO

CREATE FUNCTION GetDateString ( @val varchar(200) )
RETURNS varchar(20)
AS
BEGIN
    DECLARE @left integer;
    DECLARE @right integer;
    DECLARE @datestring varchar(20);
    DECLARE @ret varchar(20);

    set @left = PATINDEX('%200%', @val);
    set @datestring = SUBSTRING(@val, @left, 8);
    set @right = CHARINDEX('\', @datestring);
   
    set @ret = SUBSTRING(@datestring,0,@right);
    RETURN @ret;
END
GO

Alter

SET ANSI_NULLS ON
GO
SET QUOTED_IDENTIFIER ON
GO

ALTER FUNCTION [dbo].[GetDateString] ( @val varchar(200) )
RETURNS varchar(20)
AS
BEGIN
    DECLARE @left integer;
    DECLARE @right integer;
    DECLARE @datestring varchar(20);
    DECLARE @ret varchar(20);

    set @left = PATINDEX('%200%', @val);
    set @datestring = SUBSTRING(@val, @left, 12);
    set @right = CHARINDEX('\', @datestring);
   
    set @ret = SUBSTRING(@datestring,0,@right);
    RETURN @ret;
END

Test
select dbo.GetDateString('as\2009_6_22\33')

Reference

"DECLARE" http://doc.ddart.net/mssql/sql70/de-dz_2.htm
"SUBSTRING, PATINDEX, CHARINDEX" http://doc.ddart.net/mssql/sql70/fa-fz_23.htm
"CREATE FUNCTION" http://www.umachandar.com/technical/SQL2000Scripts/Main12.htm
"PATINDEX's pattern" http://www.cndw.com/tech/data/200602157855.asp
"IF" http://www.databasejournal.com/features/mssql/article.php/3087431/T-SQL-Programming-Part-1---Defining-Variables-and-IFELSE-logic.htm
"CAST" http://stackoverflow.com/questions/951320/how-to-concatenate-numbers-and-strings-to-format-numbers-in-t-sql

1) How to set Exit Code in .NET Console Application?
Use Environment.Exit method

2) How to get Exit Code in dos batch?

program.exe
If Errorlevel EQU 2 Goto C2
If Errorlevel EQU 1 Goto C1

C2:
echo "C2"
goto END

C1:
echo "C1"
goto END

END:
echo "END"

EQU - 等于

NEQ - 不等于

LSS - 小于

LEQ - 小于或等于

GTR - 大于

GEQ - 大于或等于

3) How to set Exit Code in dos batch?
SET Errorlevel=1


Reference:
http://www.allenware.com/icsw/icswidx.htm
http://hi.baidu.com/lhkyzh/blog/item/692a30ada50fdf0d4a36d6a7.html

Perl Module: File::Sort
Setup:
1. Start a command line prompt of Visual Studio
2. Run cpan in perl\bin
3. Type "install File::Sort"

Usage:
Assuming we are sorting a.tsv as below:
1       2
3       3
2       1
1       2
2       3
3       1

  use File::Sort qw(sort_file);
  sort_file({
    I => 'a.txt',
    o => 'a2.txt', k => '2n', -t => '\t'
  });

  sort_file({
    I => 'a.txt',
    o => 'a21.txt', k => ['2n','1n'], -t => '\t'
  });

Unix Sort
Examples:
sort -t, +2n
sort -t, +2nr
sort +4 -5 +1 -2 infile

sort -c
sort -t, -k2,2n -k1,1n
sort -t, -k2,2n -k1,1n -r

sort -t "`echo -e "\t"`"

Explanation:
"-t:" -> use ":" as field separator
"+2n" -> sort from the field 2,3,4... (0,1 are ignored, numbered from 0)
"+2nr" -> same as above, but reverse order
"+4 -5 +1 -2" -> sort by field [4,5), [1,2)

"-c" -> check order
"-k2n,2n -k1n,1n" -> sort by field 2,1 (others are ignored, numbered from 1)
"-r" -> reverse order

References
Perl Module
http://search.cpan.org/~cnandor/File-Sort-1.01/Sort.pm

Unix Sort
http://www.softpanorama.org/Tools/sort.shtml


BTW:
Isn't this something awesome?
#ls -l | sort +4n | awk '{print $5 "\t" $9}'

cmd /k set xx=yy&set aa=bb&c:\xxxx\xxxEnv.bat

float "1.0F"
double "1.0"
long double "1.0L"

    FILETIME now;
    GetSystemTimeAsFileTime(&now);

    ULONG uTime = now.dwHighDateTime;

Expand files recursive
compact /U /S /I

Start

perl -d abc.pl in.txt out.txt

 

View Source

"l 1-10"

"."

"v 10"

 

Step

step: "s"

step over: "n"

step out: "r"

 

Print Var

"p $name"

 

Breakpoint

set breakpoints: "b DBI::install_driver"

list breakpoints: "L"

one time breakpoint: "c 19"

delete breakpoints "B 19" "B *"

condition breakpoints "b 20 $a eq 'abc'"

 

Halt When Change

"w $count"

Delete watch: "W $count"

 

List Breakpoints and Watch

"L"

 

Stack Trace

"T"

Tool's name DiskGenius.

    unsigned __int64 u64 = 1;
    u64 = u64 << 63;
    cout << u64 << endl;
    printf("%I64u", u64 ); //right
    printf("%I64d", u64 ); //wrong

    DWORD dwVal = GetLastError();
    printf("%u", dwVal);

Scenario:
You make a data file which contains variable length records, and use another file to index from records' key hash to position in data file for random access.

Problem:
When the data file's size is larger than 2GB, ofstream.tellp() and ifstream.seekg() works wrong.
Detail (in 64-bit version):
When the file's size is between 2^31 to 2^32 , ofstream.tellp() give a negative number.
When the file's size is larger than 2^32, ofstream.tellp() give only lower 32(or may be 31) bits number.
When the file's size is larger than 2^32, ifstream.seekg() causes the ifstream invalid.

Solution:
Directly use WINAPI: CreateFile, WriteFile, ReadFile.

Scenario:
We have to use a binary executable tool, but it crashes on some input files.
We can:
1. Remove the input file which causes the tool crash
2. Remove the input files which are processed
3. Restart the tool

Problem:
When crashes the tool pop up a Just-In-Time Debugger window.  We have to manually close the window and find the problematic input file in its console output.

Analysis:
A program without debugger will forward the unhandled exception to JIT Debugger.  We can make a special "debugger" which terminates the tool program when it sees a second chance exception.
As to the file name in console output, we can use command line redirection to save the file name.  And make a special input file remover program to remove the files with names listed in the tool program's console output.

Notes for Implementation:
EnumProcesses: list processes PID
OpenProcess: PID->Process Handle
EnumProcessModules: get process' .exe name
DebugActiveProcess: as a debugger, attach to another process
Exception + FirstChance: return DBG_EXCEPTION_NOT_HANDLED
Exception + SecondChance: TerminateProcess

What Is A Thread?

1. When the thread encounter exception/interruption, its spirit is separated from its flesh.  Flesh includes stack & context data
2. IDT will may use another spirit to utilize the flesh
3. User Debugger only identifies the flesh (not the spirit)

FollowProcess <processName/processId> <next command>

For e.g.:
FollowProcess.pl excel.exe notepad.exe

Code:
my $have = 1;
while(1)
{
    open $task,"tasklist.exe|";
    $have=0;
    while(<$task>)
    {
        if( $_ =~ /$ARGV[0]/i )
        {
            $have=1;   
        }
    }
    close($task);
    if( not $have )
    {
        last;
    }
    sleep(10);
}

exec($ARGV[1]);

use File::Find;
if( scalar(@ARGV) < 2 )
{
    print "usage: MakeSingle.pl <folder> <output file>";
    exit;
}

open $output,">$ARGV[1]";
find(\&wanted, $ARGV[0]);
   


sub wanted
{
    my $file = $File::Find::name;
    if( $file =~ /.*.html$/)
    {
        print $file,"\n";
        open $fh, "<$file" or print "open failed!\n";
        my $file;   
        while(<$fh>)
        {
            $file = $file.$_;
        }
        print $file;
        $file =~ /<body>(.*)<\/body>/ims;
        print $output $1;
        close($file)
    }
}
close($output);

Alternative for big whole file reading.

        my $fileContent;
        {
          local( $/, *fh);
          open fh, "<$file" or print "open failed!\n";
          $fileContent = <fh>;
        }

Another way for file reading.
       open $fh,"<$ARGV[0]";
       @lines=<$fh>;
       close($fh);

1. GetColumn <in.tsv> <colnums 0,1,2> [remove_first_line, rfl] [use_csv, csv]
2. MergeFiles <file list> [colnames, "Name,Tel"] [use_csv, csv]
3. AppendColumn <from.tsv> <to.tsv> [use_csv, csv] [append_first_line, afl] [colname]