Reading values for variable in a file using unix - unix

I have a file parama.out having 3 variables like:
p_id
P_level
p_name
And the value for these variable will be User Input at the run-time. I tried using while loop but instead of asking for the user input its moving to the next line and throwing an error.
I used the code:
cat params.out |\
while read line
do
echo "Please provide the value for $line:\c"
read $line
done

There are a few things I used to make this work.
Put "read $line" on a seperate line (or user && or ; to seperate the two commands).
Remove the "\" after pipe to while (I removed the cat and pipe way to input the file).
Add "/dev/tty" to the read command, to let it know where the input should come from. Otherwise it'll take input from the file.
while read line
do
echo "Please provide the value for $line:"
read $line </dev/tty
echo "p_id: $p_id"
echo "p_level: $p_level"
echo "p_name: $p_name"
done < params.out
This gives the following output:
Please provide the value for p_id:
a
p_id: a
p_level:
p_name:
Please provide the value for p_level:
s
p_id: a
p_level: s
p_name:
Please provide the value for p_name:
d
p_id: a
p_level: s
p_name: d

Related

Get hour without space %time:~0,2% [duplicate]

I am compressing files using WinZip on the command line. Since we archive on a daily basis, I am trying to add date and time to these files so that a new one is auto generated every time.
I use the following to generate a file name. Copy paste it to your command line and you should see a filename with a Date and Time component.
echo Archive_%date:~-4,4%%date:~-10,2%%date:~-7,2%_%time:~0,2%%time:~3,2%%time:~6,2%.zip
Output
Archive_20111011_ 93609.zip
However, my issue is AM vs PM. The AM time stamp gives me time 9 (with a leading blank space) vs. 10 naturally taking up the two spaces.
I guess my issue will extend to the first nine days, first 9 months, etc. as well.
How do I fix this so that leading zeroes are included instead of leading blank spaces so I get Archive_20111011_093609.zip?
Another solution:
for /f "tokens=2 delims==" %%I in ('wmic os get localdatetime /format:list') do set datetime=%%I
It will give you (independent of locale settings!):
20130802203023.304000+120
( YYYYMMDDhhmmss.<milliseconds><always 000>+/-<minutes difference to UTC> )
From here, it is easy:
set datetime=%datetime:~0,8%-%datetime:~8,6%
20130802-203023
For Logan's request for the same outputformat for the "date-time modified" of a file:
for %%F in (test.txt) do set file=%%~fF
for /f "tokens=2 delims==" %%I in ('wmic datafile where name^="%file:\=\\%" get lastmodified /format:list') do set datetime=%%I
echo %datetime%
It is a bit more complicated, because it works only with full paths, wmic expects the backslashes to be doubled and the = has to be escaped (the first one. The second one is protected by surrounding quotes).
Extract the hour, look for a leading space, if found replace with a zero;
set hr=%time:~0,2%
if "%hr:~0,1%" equ " " set hr=0%hr:~1,1%
echo Archive_%date:~-4,4%%date:~-10,2%%date:~-7,2%_%hr%%time:~3,2%%time:~6,2%.zip
You should search; you can simply replace all spaces with zero set hr=%hr: =0% – jeb Oct 11 '11 at 14:16
So I did:
set hr=%time:~0,2%
set hr=%hr: =0%
Then use %hr% inside whatever string you are formatting to always get a two-digit hour.
(Jeb's comment under the most popular answer worked the best for me and is the simplest. I repost it here to make it more obvious for future users.)
As Vicky already pointed out, %DATE% and %TIME% return the current date and time using the short date and time formats that are fully (endlessly) customizable.
One user may configure its system to return Fri040811 08.03PM while another user may choose 08/04/2011 20:30.
It's a complete nightmare for a BAT programmer.
Changing the format to a firm format may fix the problem, provided you restore back the previous format before leaving the BAT file. But it may be subject to nasty race conditions and complicate recovery in cancelled BAT files.
Fortunately, there is an alternative.
You may use WMIC, instead. WMIC Path Win32_LocalTime Get Day,Hour,Minute,Month,Second,Year /Format:table returns the date and time in a invariable way. Very convenient to directly parse it with a FOR /F command.
So, putting the pieces together, try this as a starting point...
SETLOCAL enabledelayedexpansion
FOR /F "skip=1 tokens=1-6" %%A IN ('WMIC Path Win32_LocalTime Get Day^,Hour^,Minute^,Month^,Second^,Year /Format:table') DO (
SET /A FD=%%F*1000000+%%D*100+%%A
SET /A FT=10000+%%B*100+%%C
SET FT=!FT:~-4!
ECHO Archive_!FD!_!FT!.zip
)
I found the best solution for me, after reading all your answers:
set t=%date%_%time%
set d=%t:~10,4%%t:~7,2%%t:~4,2%_%t:~15,2%%t:~18,2%%t:~21,2%
echo hello>"Archive_%d%"
If AM I get 20160915_ 150101 (with a leading space and time).
If PM I get 20160915_2150101.
#For /F "tokens=1,2,3,4 delims=/ " %%A in ('Date /t') do #(
Set DayW=%%A
Set Day=%%B
Set Month=%%C
Set Year=%%D
Set All=%%D%%B%%C
)
"C:\Windows\CWBZIP.EXE" "c:\transfer\ziptest%All%.zip" "C:\transfer\MB5L.txt"
This takes MB5L.txt and compresses it to ziptest20120204.zip if run on 4 Feb 2012
You can add leading zeroes to a variable (value up to 99) like this in batch:
IF 1%Var% LSS 100 SET Var=0%Var%
So you'd need to parse your date and time components out into separate variables, treat them all like this, then concatenate them back together to create the file name.
However, your underlying method for parsing date and time is dependent on system locale settings. If you're happy for your code not to be portable to other machines, that's probably fine, but if you expect it to work in different international contexts then you'll need a different approach, for example by reading out the registry settings:
HKEY_CURRENT_USER\Control Panel\International\iDate
HKEY_CURRENT_USER\Control Panel\International\iTime
HKEY_CURRENT_USER\Control Panel\International\iTLZero
(That last one controls whether there is a leading zero on times, but not dates as far as I know).
From the answer above, I have made a ready-to-use function.
Validated with french local settings.
:::::::: PROGRAM ::::::::::
call:genname "my file 1.txt"
echo "%newname%"
call:genname "my file 2.doc"
echo "%newname%"
echo.&pause&goto:eof
:::::::: FUNCTIONS :::::::::
:genname
set d1=%date:~-4,4%
set d2=%date:~-10,2%
set d3=%date:~-7,2%
set t1=%time:~0,2%
::if "%t1:~0,1%" equ " " set t1=0%t1:~1,1%
set t1=%t1: =0%
set t2=%time:~3,2%
set t3=%time:~6,2%
set filename=%~1
set newname=%d1%%d2%%d3%_%t1%%t2%%t3%-%filename%
goto:eof
As others have already pointed out, the date and time formats of %DATE% and %TIME% (as well as date /T and time /T) are locale-dependent, so extracting the current date and time is always a nightmare, and it is impossible to get a solution that works with all possible formats since there are hardly any format limitations.
But there is another problem with a code like the following one (let us assume a date format like MM/DD/YYYY and a 12 h time format like h:mm:ss.ff ap where ap is either AM or PM and ff are fractional seconds):
rem // Resolve AM/PM time:
set "HOUR=%TIME:~,2%"
if "%TIME:~-2%" == "PM" if %HOUR% lss 12 set /A "HOUR+=12"
if "%TIME:~-2%" == "AM" if %HOUR% equ 12 set /A "HOUR-=12"
rem // Left-zero-pad hour:
set "HOUR=0%HOUR%"
rem // Build and display date/time string:
echo %DATE:~-4,4%%DATE:~0,2%%DATE:~3,2%_%HOUR:~-2%%TIME:~3,2%%TIME:~6,2%
Each instance of %DATE% and %TIME% returns the date or time value present at the time of its expansion, therefore the first %DATE% or %TIME% expression might return a different value than the following ones (you can prove that when echoing a long string containing a huge amount of such, preferrably %TIME%, expressions).
You could improve the aforementioned code to hold a single instance of %DATE% and %TIME% like this:
rem // Store current date and time once in the same line:
set "CURRDATE=%DATE%" & set "CURRTIME=%TIME%"
rem // Resolve AM/PM time:
set "HOUR=%CURRTIME:~,2%"
if "%CURRTIME:~-2%" == "PM" if %HOUR% lss 12 set /A "HOUR+=12"
if "%CURRTIME:~-2%" == "AM" if %HOUR% equ 12 set /A "HOUR-=12"
rem // Left-zero-pad hour:
set "HOUR=0%HOUR%"
rem // Build and display date/time string:
echo %CURRDATE:~-4,4%%CURRDATE:~0,2%%CURRDATE:~3,2%_%HOUR:~-2%%CURRTIME:~3,2%%CURRTIME:~6,2%
But still, the returned values in %DATE% and %TIME% could reflect different days when executed at midnight.
The only way to have the same day in %CURRDATE% and %CURRTIME% is this:
rem // Store current date and time once in the same line:
set "CURRDATE=%DATE%" & set "CURRTIME=%TIME%"
rem // Resolve AM/PM time:
set "HOUR=%CURRTIME:~,2%"
if "%CURRTIME:~-2%" == "PM" if %HOUR% lss 12 set /A "HOUR+=12"
if "%CURRTIME:~-2%" == "AM" if %HOUR% equ 12 set /A "HOUR-=12"
rem // Fix date/time midnight discrepancy:
if not "%CURRDATE%" == "%DATE%" if %CURRTIME:~0,2% equ 0 set "CURRDATE=%DATE%"
rem // Left-zero-pad hour:
set "HOUR=0%HOUR%"
rem // Build and display date/time string:
echo %CURRDATE:~-4,4%%CURRDATE:~0,2%%CURRDATE:~3,2%_%HOUR:~-2%%CURRTIME:~3,2%%CURRTIME:~6,2%
Of course the occurrence of the described problem is quite improbable, but at one point it will happen and cause strange unexplainable failures.
The described problem cannot occur with the approaches based on the wmic command as described in the answer by user Stephan and in the answer by user PA., so I strongly recommend to go for one of them. The only disadvantage of wmic is that it is way slower.
Your question seems to be solved, but ...
I'm not sure if you take the right solution for your problem.
I suppose you try to compress each day the actual project code.
It's possible with ZIP and 1980 this was a good solution, but today you should use a repository system, like subversion or git or ..., but not a zip-file.
Ok, perhaps it could be that I'm wrong.
I realise this is a moot question to the OP, but I just brewed this, and I'm a tad proud of myself for thinking outside the box.
Download gawk for Windows at http://gnuwin32.sourceforge.net/packages/gawk.htm .... Then it's a one liner, without all that clunky DOS batch syntax, where it takes six FOR loops to split the strings (WTF? That's really really BAD MAD AND SAD! ... IMHO of course)
If you already know C, C++, Perl, or Ruby then picking-up AWK (which inherits from the former two, and contributes significantly to the latter two) is a piece of the proverbial CAKE!!!
The DOS Batch command:
echo %DATE% %TIME% && echo %DATE% %TIME% | gawk -F"[ /:.]" "{printf(""""%s%02d%02d-%02d%02d%02d\n"""", $4, $3, $2, $5, $6, $7);}"
Prints:
Tue 04/09/2012 10:40:38.25
20120904-104038
Now that's not quite the full story... I'm just going to be lazy and hard-code the rest of my log-file-name in the printf statement, because it's simple... But if anybody knows how to set a %NOW% variable to AWK's output (yeilding the guts of a "generic" now function) then I'm all ears.
EDIT:
A quick search on Stack Overflow filled in that last piece of the puzzle, Batch equivalent of Bash backticks.
So, these three lines of DOS batch:
echo %DATE% %TIME% | awk -F"[ /:.]" "{printf(""""%s%02d%02d-%02d%02d%02d\n"""", $4, $3, $2, $5, $6, $7);}" >%temp%\now.txt
set /p now=<%temp%\now.txt
echo %now%
Produce:
20120904-114434
So now I can include a datetime in the name of the log-file produced by my SQL Server installation (2005+) script thus:
sqlcmd -S .\SQLEXPRESS -d MyDb -e -i MyTSqlCommands.sql >MyTSqlCommands.sql.%now%.log
And I'm a happy camper again (except life was still SOOOOO much easier on Unix).
I prever to use this over the current accepted answer from Stephan as it makes it possible to configure the timestamp using named parameters after that:
for /f %%x in ('wmic path win32_utctime get /format:list ^| findstr "="') do set %%x
It will provide the following parameters:
Day
DayOfWeek
Hour
Milliseconds
Minute
Month
Quarter
Second
WeekInMonth
Year
You can then configure your format like so:
SET DATE=%Year%%Month%%Day%
So you want to generate date in format YYYYMMDD_hhmmss.
As %date% and %time% formats are locale dependant you might need more robust ways to get a formatted date.
Here's one option:
#if (#X)==(#Y) #end /*
#cscript //E:JScript //nologo "%~f0"
#exit /b %errorlevel%
#end*/
var todayDate = new Date();
todayDate = "" +
todayDate.getFullYear() +
("0" + (todayDate.getMonth() + 1)).slice(-2) +
("0" + todayDate.getDate()).slice(-2) +
"_" +
("0" + todayDate.getHours()).slice(-2) +
("0" + todayDate.getMinutes()).slice(-2) +
("0" + todayDate.getSeconds()).slice(-2) ;
WScript.Echo(todayDate);
and if you save the script as jsdate.bat you can assign it as a value :
for /f %%a in ('jsdate.bat') do #set "fdate=%%a"
echo %fdate%
or directly from command prompt:
for /f %a in ('jsdate.bat') do #set "fdate=%a"
Or you can use powershell which probably is the way that requires the less code:
for /f %%# in ('powershell Get-Date -Format "yyyyMMdd_HHmmss"') do set "fdate=%%#"
Adding other options to this list of answers.
you could have replaced empty space with a 0 something like echo %time: =0%
but that is still dependent, move that code to a buddy's PC in some other random place and you'll get funny outputs. So you can incorporate powershell's Get-Date:
for /f "tokens=*" %%i in ('PowerShell -Command "Get-Date -format 'yyyymmdd_HHmmss'"') do echo %%i.zip"
A space is legal in file names. If you put your path and file name in quotes, it may just fly. Here's what I'm using in a batch file:
svnadmin hotcopy "C:\SourcePath\Folder" "f:\DestPath\Folder%filename%"
It doesn't matter if there are spaces in %filename%.

reading from dynamically set filenames in R

I am writing a CGI script in Perl with a section of embedded R script which produces a graph. The original data filename is unknown as it has been uploaded by the CGI script and is stored in a Perl variable called $filename.
My question is that I now would like to open that file in R using read.table(). I am using Statistics::R and so I have tried:
my $R = Statistics::R->new();
$R->set('filename',$filename);
my $out1 = $R->run(
q`rm(list=ls())`,
# Fetch data
q`setwd("/var/www/uploads")`,
q`peakdata<-read.table(filename, sep="",col.names=c("mz","intensity","ionsscore","matched","query","index","hit"))`,
q`attach(peakdata)` ...etc
I can get this to work ONLY if I change $filename into something static and known like 'data.txt' before trying to open the file in read.table - is there a way for me to open a file with a variable for a name?
Thank you in advance.
One possible way to do this is by doing a little more work in Perl.
This is untested code, to give you some ideas:
my $filename = 'fileNameIGotFromSomewhere.txt'
my $source_dir = '/var/www/uploads';
my $file = "$source_dir/$fielname";
# make sure we can read it
unless ( -r $file ) {
die 'can read that data file: $!";
}
Then instead of $R->set, you could interpolate the file name into the R program. Where you've used the single-quote operator, use the double-quote operator instead:
So instead of:
q`peakdata<-read.table(filename, sep="",col.names= .... )`
Use:
qq`peakdata<-read.table($filename, sep="",col.names= .... )`
Now this looks like it would be inviting problems similar to SQL/Code Injections, so that's why I put int the logic to insure that the file exists and is readable. You might be able to think other checks to add to safeguard your use of user-supplied info.

PROGRESS - Validating a user-input file output path

I've written some PROGRESS code that outputs some data to a user defined file. The data itself isn't important, the output process works fine. It's basically
DEFINE VARIABLE filePath.
UPDATE filePath /*User types in something like C:\UserAccount\New.txt */
OUTPUT TO (VALUE) filePath.
Which works fine, a txt file is created in the input directory. My question is:
Does progress have any functionality that would allow me to check if an input
file path is valid? (Specifically, if the user has input a valid directory, and if they have permission to create a file in the directory they've chosen)
Any input or feedback would be appreciated.
FILE-INFO
Using the system handle FILE-INFO gives you a lot of information. It also works on directories.
FILE-INFO:FILE-NAME = "c:\temp\test.p".
DISPLAY
FILE-INFO:FILE-NAME
FILE-INFO:FILE-CREATE-DATE
FILE-INFO:FILE-MOD-DATE
FILE-INFO:FILE-INFO
FILE-INFO:FILE-MOD-TIME
FILE-INFO:FILE-SIZE
FILE-NAME:FILE-TYPE
FILE-INFO:FULL-PATHNAME
WITH FRAME f1 1 COLUMN SIDE-LABELS.
A simple check for existing directory with write rights could be something like:
FUNCTION dirOK RETURNS LOGICAL (INPUT pcDir AS CHARACTER):
FILE-INFO:FILE-NAME = pcDir.
IF INDEX(FILE-INFO:FILE-TYPE, "D") > 0
AND INDEX(FILE-INFO:FILE-TYPE, "W") > 0 THEN
RETURN TRUE.
ELSE
RETURN FALSE.
END FUNCTION.
FILE-NAME:FILE-TYPE will start with a D for directories and a F for plain files. It also includes information about reading and writing rights. Check the help for more info. If the file doesn't exist basically all attributes except FILE-NAME will be empty or unknown (?).
Edit: it seems that FILE-TYPE returns W in some cases even if there's no actual writing rights in that directory so I you might need to handle this through error processing instead
ERROR PROCESSING
OUTPUT TO VALUE("f:\personal\test.txt").
PUT UNFORMATTED "Test" SKIP.
OUTPUT CLOSE.
CATCH eAnyError AS Progress.Lang.ERROR:
/* Here you could check for specifically error no 98 indicating a problem opening the file */
MESSAGE
"Error message and number retrieved from error object..."
eAnyError:GetMessage(1)
eAnyError:GetMessageNum(1) VIEW-AS ALERT-BOX BUTTONS OK.
END CATCH.
FINALLY:
END FINALLY.
SEARCH
When checking for a single file the SEARCH command will work. If the file exists it returns the complete path. It does however not work on directory, only files. If you SEARCH without complete path e g SEARCH("test.p") the command will search through the directories set in the PROPATH environment variable and return the first matching entry with complete path. If there's no match it will return unknown value (?).
Syntax:
IF SEARCH("c:\temp\test.p") = ? THEN
MESSAGE "No such file" VIEW-AS ALERT-BOX ERROR.
ELSE
MESSAGE "OK" VIEW-AS ALERT-BOX INFORMATION.
SYSTEM-DIALOG GET-FILE character-field has an option MUST-EXIST if you want to use a dailogue to get filename/dir from user. Example from manual
DEFINE VARIABLE procname AS CHARACTER NO-UNDO.
DEFINE VARIABLE OKpressed AS LOGICAL INITIAL TRUE.
Main:
REPEAT:
SYSTEM-DIALOG GET-FILE procname
TITLE "Choose Procedure to Run ..."
FILTERS "Source Files (*.p)" "*.p",
"R-code Files (*.r)" "*.r"
MUST-EXIST
USE-FILENAME
UPDATE OKpressed.
IF OKpressed = TRUE THEN
RUN VALUE(procname).
ELSE
LEAVE Main.
END.

Using Variables in Declared in Different File on Unix Script

How can I keep variables used in a script in a secondary file? For example, use store variables in myscript.env for use in a script myscript.sh.
So whenever I need to change variables I can edit myscript.env instead of myscript.sh.
Simply include the file using the "dot operator".
assuming these two files are in the same directory, the following will work:
t.sh:
#!/usr/bin/sh
# Note the first dot on the following line
. ./t.env
echo $TESTVAR
t.env:
TESTVAR="Hello world"
When run:
~/tmp$ sh ./t.sh
Hello world
This is quite a usual design pattern. Think along the lines of
myscript.sh:
#!/path/to/shell
CONFIG_FILE=myscript.env
CONFIG_DIR=`dirname $0`
# or e.g. CONFIG_DIR=/etc/myscript
CONFIG="$CONFIG_DIR/$CONFIG_FILE"
. $CONFIG
echo "SOMEVAR=$SOMEVAR"
myscript.env
SOMEVAR="The value of some var"
Now /path/to/myscript.sh will output SOMEVAR=The value of some var

Unix C Shell Scripting printf help

Attempting to print out a list of values from 2 different variables that are aligned correctly.
foreach finalList ($correctList $wrongList)
printf "%20s%s\n" $finalList
end
This prints them out an they are aligned, but it's one after another. How would I have it go through each item in each list and THEN go to a new line?
I want them to eventually appear like this:
Correct Incorrect
Good1 Bad1
Good2 Bad2
Good3 Bad3
Good comes from correctList
Bad comes from wrongList
Getting rid of \n makes it Like this:
Good1 Bad1 Good2 Bad2
I just want 2 columns.
You can iterate over both lists at the same time like this:
# Get the max index of the smallest list
set maxIndex = $#correctList
if ( $#wrongList < $#correctList ) then
set maxIndex = $#wrongList
endif
set index = 1
while ($index <= $maxIndex)
printf "%-20s %s\n" "$correctList[$index]" "$wrongList[$index]"
# index++
end
try getting rid of the \n
I believe the pr(1) command with the -m option will help do what you want. Look at its man page to eliminate the header/trailer options and set the column widths.
Also, I recommend you not use the C-Shell for scripting; you'll find the sh-syntax shells (sh, bash, ksh, etc) are more consistent and much easier to debug.

Resources