Search and Replace Text in Multiple Files in Only 7 Lines

Have you ever needed to perform search / replace operations on multiple files in a folder and all its subfolders? Have you ever needed to search / replace text in a list of files? How about carrying out multiple search / replace operations on a single file based on search / replace pairs specified in a text file? This article shows how to do each of these common search / replace operations using TEXTools--and believe it or not--in only 7 lines of batch code!To carry out each of these tasks, we'll be utilizing TEXTools (TCL.exe) from a batch file. TEXTools is the perfect compliment to batch file programming as it greatly extends the power of batch files. Without implementing a single for loop, a TEXTools endowed batch file can nonetheless easily process multiple files. This is made possible because batch files can be constructed on-the-fly by another batch file and then called by that batch file, (its kind of like a train that builds its track as it goes). The reason that a for loop isn't needed here to process multiple files is simply because TEXTools can take a list of files that need to be processed and translate that list into a batch file that, when run, results in the processing of each of the files. For example, suppose we wanted to carry out some relatively complex text editing operations on the following list of files:myfile.txtyourfile.txttheirfile.txt...One way to do this would be to place the editing logic into a batch file called procfile.bat and then process each file--one after the other--using procfile.bat like this:procfile.bat myfile.txtprocfile.bat yourfile.txtprocfile.bat theirfile.txt...That's great if you only have 3 files, but what if you have 100 or even 1000 files to process? That's where TEXTools comes into the picture. If batch files are just text and they can be constructed on-the-fly, then why not use TEXTools to turn our list of filenames into a batch file that, when run, results in the processing of each of the files? If a batch file like procfile.bat is going to be employed in the process anyway, then we're already half-way to our goal. What we want is for the list of file names to be preceded by "procfile.bat" with one change--procfile.bat needs to be called. Now we have a batch file that will do all the work for us:call procfile.bat myfile.txtcall procfile.bat yourfile.txtcall procfile.bat theirfile.txt...Of course, this is the kind of thing that TEXTools excels at doing. Adding the text, "call procfile.bat " to the front of every line in a text file containing hundreds of file names--in order to create and execute an on-the-fly batch file is a snap:type ListOfFiles.txt | tcl InsStr 1 'call procfile.bat ' >t$1.batcall t$1.batWe're now ready to tackle our assignment except for one missing puzzle piece. We need our procfile.bat. Ours needs to take as arguments a file name, a search text and a replacement text:ProcFile Internally, ProcFile.bat will edit the given file by replacing all occurrences of with inside the file. Fundamentally, this could be done with just two lines of code (under Windows 95 and 98, etc. this could be done in a single line without the need of a temporary intermediate file):type %1 | tcl "ReplStr '%2' '%3'" >t$1.txttype t$1.txt >%1In reality, we need a bit more than this though. What if the search text or replacement text is specified having embedded blanks? This can only be done at the command prompt by using surrounding double-quotes like so:ProcFile myfile.txt "red car" "blue car"The problem is that these surrounding double-quotes are part of the parameter as far as Windows is concerned. As far as we're concerned however, they're not and so we must get rid of them. Again, this is a simple matter using TEXTools:echo %2| tcl "ReplStr '#22' '' | InsStr 1 'set search='" >t$1.batecho %3| tcl "ReplStr '#22' '' | InsStr 1 'set replace='" >>t$1.batcall t$1.batThese three magical lines simply remove the surrounding double-quotes from each of the search / replace parameters that are passed into ProcFile. Not that the parameters themselves are altered in any way. Their unquoted contents are merely copied into environment variables! Note again the use of an on-the-fly batch file to accomplish this. Laying down tracks...So now, our ProcFile.bat looks something like this::: Remove any surrounding double-quotes from the search/replace text...echo %2| tcl "ReplStr '#22' '' | InsStr 1 'set search='" >t$1.batecho %3| tcl "ReplStr '#22' '' | InsStr 1 'set replace='" >>t$1.batcall t$1.bat:: Perform text replacements on this file...type %1 | tcl "ReplStr '%search%' '%replace%'" >t$1.txttype t$1.txt >%1Bear in-mind that this is a bare-bones implementation of ProcFile.bat. Normally, you'd include some logic to abort execution if all parameters weren't supplied, etc. but this is just a demonstration. For the actual batch files, see the "Search and Replace" examples available in the TEXTools download package (installed beneath the "Sample Scripts" folder) and also available online at www.fireflysoftware.comTEXToolsexamples.htm. Now, on to bigger and better things...Using a Recursive File SearchNow that we have a batch file that performs text replacements on a single file, how can we employ it to process all files in a given folder and its subfolders? The key to this lies in the DIR command. When used with the /b and /s switches, the DIR command lists (fully qualified) all files in a folder (and all its subfolders) that match a given file specification. Imagine that... a list of all the files in a given folder and its subfolders that match say, the file specification, "work*.txt"! In practical terms that translates into a gem like this:dir %1 /b /s | tcl "InsStr 1 'call ProcFile.bat #22' | AppendStr '#22 %2 %3'" >t$2.batcall t$2.batHere's a batch file (say, "Recursive.bat") that can be called as follows:Recursive work*.txt "this" "that"Here, it would replace all instances of "this" with "that" in all files (in the current folder and subfolders) matching "work*.txt". Not bad for 7 lines of batch code, eh?Using a List of FilesIf the files to be edited are listed in a text file, the batch code needed to process them is almost identical to the above:type %1 | tcl "InsStr 1 'call ProcFile.bat #22' | AppendStr '#22 %2 %3'" >t$2.batcall t$2.batThe only difference is in where the list of files originates. With this approach, all you need is a list of files in a text file to process. Such a batch file might be called like this:ProcFileList FileList.txt "this" "that"This would replace all instances of "this" with "that" in all of the files listed in FileList.txt.Using a List of ReplacementsThis last example of search and replace performs multiple edits on a single file by utilizing a file containing multiple search / replace pairs. Again, a call to ProcFile is all that is needed:type %2 | tcl "InsStr 1 'call ProcFile.bat %1 '" >t$2.batcall t$2.batSuch a batch file as this might be called ProcReplList.bat. Here's how it would be called to perform multiple search / replacement operations on MyFile.txt using the search / replace pairs found in Replacements.txt:ProcReplList MyFile.txt Replacements.txtNote: Each of the above search / replace examples can be found online at www.fireflysoftware.comTEXToolsexamples.htm or inside the TEXTools download package, (installed beneath the "Sample Scripts" folder).Permission is granted to freely post or distribute this article so long as it is unedited and credit and copyright information is left intact.http://www.fireflysoftware.com/rdsearchreplace.htmCopyright © 2005 Firefly SoftwareBarry Block Author of TEXTools, a powerful text editing workbench for Windows http://www.fireflysoftware.com/rdsearchreplace.htm

Barry Block