I have a folder into which a number of MSQL scripts get dropped into after each weekly sprint. For example, 10 scripts were placed into the folder today. I had to then open each script individually and run it against the applicable database. The database that it needs to be run against is in the name of the file.
e.g. [2] [CRMdata]UpdateProc.sql
The [2] represents the sequence in which it is run, so script [1] needs to be run before it.
[CRMdata] is the database I have to run it against.
This process is very tiresome, especially if there are 50 scripts to run sequentially.
I was wondering if there was an easier way to do this?
Perhpas a .bat file, which reads the filename, and executes the scripts sequentially based on the script number, as well as executing it against the database specified in the file name.
Any help is much appreciated.
Thanks.
First, when you need to run things, consider using SQL Server Job Agent. This is a good way to schedule simple things.
For a task like this, I would recommend PowerShell in combination with "sqlcmd". This command is actually the answer to your question, since it will run scripts from the command line.
However, go a step further. Schedule a job that runs once per week (or whenever you want it run). Have it consist of one step, a PowerShell script. This can then loop through all the scripts in the directory, extract the file name from the name, and run the script using sqlcmd. Along the way, also log what you are doing in a table so you can spot errors.
I don't know anything about executing SQL with MSQL. You will have to work out how to run each script against the proper database using whatever command-line utility is provided for MSQL.
I can help you with a batch file that will sort the SQL files in the correct sequence order, and parse out the name of the database.
The job is much easier in batch if the sequence numbers are zero prefixed to be a constant width. I'm assuming it is OK to rename the files, so that is what this solution does.
I also assumed you will never have more than 999 files to process. The code can easily be modified to handle more.
Some changes will have to be made if any file names contain the ! character because delayed expansion will corrupt the expansion of the FOR variables. But that is an unlikely problem.
#echo off
setlocal enableDelayedExpansion
:: Change the definition to point to the folder that contains the scripts
set "folder=sqlCodeFolder"
:: The mask will only match the pattern that you indicated in your question
set "mask=[*] [*]*.sql"
:: Rename the .sql files so that the sequence numbers are zero prefixed
:: to width of 3. This enables the default alpha sort of the directory to be
:: in the proper sequence
for /f "tokens=1* delims=[]" %%A in ('dir /b "%folder%\%mask%"') do (
set seq=00%%A
ren "%folder%\[%%A]%%B" "[!seq:~-3!]%%B"
)
::Process the renamed files in order
for %%F in ("%folder%\%mask%") do (
for /f "tokens=2 delims=[] " %%D in ("%%~nF") do (
rem %%F contains the full path to the sql file
rem %%D contains the name of the database, without enclosing []
rem Replace the echo line below with the proper command to run your script
echo run %%F against database [%%D]
)
)
Related
I am very new to creating batch files.
I have to run a command, with an increasing file number e.g
c:>program.bat -propertyfile "1.property"
Right now, I have to type the command manually, wait 1 minute, then type the command again by increasing the property file # i.e "2.property" "3.property" "4.property" etc....
I want to automate this, and still would like to see the results in the command prompt as it runs.
How can this be accomplished?
See https://ss64.com/nt/for.html and specifically https://ss64.com/nt/for_l.html
FOR /L %%G IN (1,1,4) DO program.bat -propertyfile "%%G.property"
Should run your command for files 1.property to 4.property but if you're actually running it for files in a directory rather than a list of integers one of the other FOR constructs might be more appropriate. Perhaps https://ss64.com/nt/for_r.html
This question is probably unnecessary as I probably have been searching for the wrong criteria. It's pretty easy I think but I just can't find the answer. I use batch files to automate installs with my work, setting folder permissions, copying files here, and there, silently running programs, removing old ones etc. The batch files contain lots of repetition and I want to tidy them up greatly for easier management. I have decided to set the repeated commands and copy folder locations as variables then use them instead. This is all fine but I'm adding the variables as a LIST which looks rubbish to me, I'm sure I can load them together in a sentence rather than a new line for each. Here's what I mean...
set dir1=md c:\newfolder
set killtask=taskkill /im someprog.exe /f >nul 2>&1
set config=echo F| XCOPY %~dp0configfile3.cfg /y C:\newfolder
And on and on...
So at the minute my batch now looks like this
JOB1
%dir1%
%killtask%
%config%
What I want to be able to do is have it like this (to reduce length if batch file etc.)
JOB1
%dir1% %killtask% %config%
(summary of my comments to the question):
Check your variables! Some chars like pipe (|) or redirection (>) change how the line is interpreted. Use this syntax to correct it:
set "killtask=taskkill /im someprog.exe /f >nul 2>&1"
(note the quotes and their position)
you can call several commands with &:
echo hello&echo world.
With your complex variables this would probably fail/work in an unexpected way. Try
(%dir1%)&(%killtask%)&(%config%)
so that the redirections and pipes stay at their intended commands.
I'm having trouble with a kinda specific problem.
The monitoring software (used for robots in the manufacturing halls) used by the company i am working for, generates a log file (.sdat) every 15 minutes. The content of a log file looks like this:
The syntax: Time;machine;status
13:53:23;KP85;ms:9999
13:53:49;KP85;ms:3
13:54:54;KP85;ms:4
14:06:04;KP85;ms:9999
13:51:38;Robot1;ms:9999
etc...
I've managed to concatenate all the log files into one big file, including the filename at the start of each new row, like this:
The syntax: Filename:Time;Machine;Status
01-03-2016-00-20.sdat:0:07:40;KP65;ms:3
01-03-2016-00-20.sdat:0:09:09;KP65;ms:4
01-03-2016-00-20.sdat:0:09:11;KP65;ms:3
01-03-2016-00-20.sdat:0:09:13;KP65;ms:4
etc..
The reason I did this is because i need the time as well as the date(which is included in the filename) a certain status for a machine has been logged. However, if i import this into SQL management studio, it recognizes 3 columns instead of 4, because the filename and first column(time) are separated by a ':' instead of a ';'. I tried solving it with an SQL Query, separating the date and time with a LEFT() and RIGHT(), but guess what: the time field format changes when the time switches from 9:59:59 to 10:00:00, creating an extra character for the time field (so the data in the column would look like this ':9:59:59', which isn't a valid time field). Perhaps it could be done with SQL but it just seems to me like it would take too much complexity in the SQL code for such a small problem.
So at this point, i thought it would be better to tackle this problem early on; within the batch file which generates the large file, so Management Studio does recognize 4 instead of 3 columns. This is how my .bat file looks like at the moment:
#echo off
findstr "^" *.sdat >output.txt
What do i have to do to get this right?
Thanks in advance,
Mike Sohns
#echo off
(for %%f in (*.sdat) do (
for /f "delims=" %%a in (%%f) do #echo %%f;%%a
))>output.txt
NOTE: this will swallow empty lines, but I think in your case that's not a problem. If yes, the code can be adapted.
Given that all the log files you are concatenating have the sdat extension, you have to replace sdat: for sdat; in order to have a ; separated CSV.
To achieve this, you can use the batch script in this other answer(How to replace substrings in windows batch file), that replaces substrings in a text file using a batch script.
Sorry not much experience with batch files hence help needed please! ;-)
I'm working in a DOS box on a Windows 7 64 bit system.
I want to run an application as a batch file, but reading the information it needs from a text file which can be updated/amended regularly.
The syntax of the basic application is:
appname "variable" (the variable MUST be enclosed in quotes)
Successive variables can be concatenated to the following single line format:
appname "var1" "var2" "var3" "var4" ... etc
So I've created a batch file containing the above. However, this is unweildy when it comes to updating. Sometimes I omit the delimiting quotes which creates problems in the execution of the batch file.
It seems to me that from an updating/amending point it would be easier to set up a text file, say text.txt which would contain the following information:
"var1"
"var2"
"var3"
"var4"
etc. on successive lines.
This would make it easier for me to update and also to ensure I don't omit the delimiting quotes.
The batch file would get the application to "read" the text file, execute the first variable, then the second etc all the way through to the end. But I'm not sure if this is possible and if so, how to get the batch file to read successive lines in the text file and use those variables.
As I said earlier, I've not much experience with batch files and don't have a clue how to do this! :-(
Help please, thanks
Alan
Like this :
#echo off
set $textFile="test.txt"
for /f "delims=" %%a in ('type %$textFile%') do appname.exe %%a
#ECHO OFF
SETLOCAL ENABLEDELAYEDEXPANSION
SET "batchline=appname"
FOR /f "delims=" %%a IN (q25193799.txt) DO SET "batchline=!batchline! "%%~a""
ECHO(%batchline%
GOTO :EOF
I used a file named q25193799.txt containing your data for my testing.
The application line is merely echoed for verification. Remove the ECHO( after verification to execute your application.
The data in the file need not have "enclosing quotes".
Thanks to both responses, both different but good in their own way. I use a few different instances and I've decided to use the first response for the longer lists and the second one for shorter lists (that way I can check see if something is going wrong, because I might have missed the second delimiting quote in the text file.
Superb, thanks to you both.
I have a section of code;
:check
for /f "tokens=1* delims==" %%A in ('"wmic process get description, commandline /format:list"') do (
if "%%A"=="CommandLine" (
set "cmd=%%B"
) else if "%%A"=="Description" (
set "desc=%%B"
setlocal enableDelayedExpansion
set "desc=!desc:~0,-1!"
set "cmd=!cmd:~0,-1!"
if /i !desc! == %1 (
echo !cmd! >>C:\test.txt
)
endlocal
)
)
goto:eof
Which pretty much works (this is actually a function called from withing a batch file e.g
call:check processname1.exe
call:check processname2.exe
call:check processname3.exe
etc...
What I'd like to do (if possible), is, insead of echoing to a file, I'd like to be able to create 2 variables. something like;
processname1.exe processname3.exe <-- (for each process 'checked' if it IS running, append its name to this variable)
commandlinepath1 commandlinepath2 <-- (for each process 'checked' if it IS running, append its path to this variable)
If this is possible, and I can then call on these variable later in my script, I'd like to be able to tskill the running processes (easy enough if the variables above can be made), then later on, RE-OPEN these processes (using and command line parameter that were in the original path. This is where I'm lost.
My code above (writing to a file). will give results like;
"C:\somefolder\someexe.exe" -some_parameter
"C:\some therfolder\someotherexe.exe"
"C:\another older\anotherexe.exe" param1 param2
But What I need to do, is take each line of this file (or variable if possible), and run them (if I copy each line into the RUN command of windows, it works, but doing it through CMD it doesn't).
I've tried using a for loop to open the files, and it does, except the script waits for the process to finish beford continuing (and these process won't end, since they are applications). If I try to use START .. then it loads a new CMD window??
What I need to do (in case there is a better option) is
for a pre-determined set of processes, check to see if they are running
kill the ones that are (if they are not, fine ignore it)
delete some files (I can do this, the reason for killin the processes is they hold the files open, preventing deletion)
Re-open all the programs that were originally running
Thanks..
Not a direct answer, but since you already use wmic, maybe using its built in capabilities (query, start & stop) would make your goal easier to achieve?
I come up with the following:
#echo off
setlocal
for %%C in (notepad.exe) do (
for /f "skip=1 tokens=2 delims==" %%F in ('wmic process where description^="%%C" get commandline /format:list') do (
REM required to normalize unicode output from WMIC
set commandline=%%F
REM '\=\\' required as wmic treats \ as escape char in query
call wmic process where commandline='%%commandline:\=\\%%' terminate
REM do your work here
call wmic process call create '%%commandline%%'
)
)
What it does:
First for supplies process names. In my example, it simply is notepad.exe, but you could call with a list: for (process1 process2 process3), or replace it with for /f to supply values from file. If you want to use quoted names, you would have to remove quote from next line (description^="%%C").
Second for does real work: it gets a list of all processes matching description and sequentially stops and starts each of them.
To try it, simply put it in a batfile.bat, open notepad(s) and execute.
Note: if you open notepad with a file, either specify an absolute path, or do it via explorer (double click). The issue here is of current directory - which you could also stumble upon if any of your processes does reference relative paths (unlikely, but not impossible)
Last but not least - doing that in powershell would be the easiest, shortest and most reliable.