If you happen to have the JREPL.BAT regular expression text processing utility (v7.9 or later), then you can use:
jrepl "," ";" /p "([\c\q]+)|\q.*?\q" /prepl "$1?{$0}:$0" /f "test.csv" /o -
Use call jrepl
if you put the command within a batch script.
The original file will be overwritten. You can substitute a new file name for -
if you don't want to overwrite the original.
JREPL.BAT is pure script (hybrid JScript/batch) that runs natively on any Windows machine from XP onward - no 3rd party .exe file required.
The JREPL solution works by performing the replacement in two steps.
1) The /P
option breaks each line into unquoted strings and quoted strings. The /PREPL
option passes unquoted strings on to the normal FIND/REPLACE, and unquoted strings are preserved as is.
2) The main FIND/REPLACE substitutes ;
for ,
It is possible to reliably accomplish this with pure batch using a variant of a technique developed by jeb at 'Pretty print' windows %PATH% variable - how to split on ';' in CMD shell. Although any pure batch solution will be significantly slower than hybrid solutions like JREPL.BAT, ParseCSV.bat, or a powershell solution.
Here is a batch script derived from jeb's technique - simply pass the name of the CSV file as the one and only argument. The original file will be overwritten. It should be trivial to modify the script to write the output to a new file instead. See jeb's post for an overview of how this seemingly magical technique works.
@echo off
setlocal disableDelayedExpansion
>"%~1.new" (
for /f usebackq^ delims^=^ eol^= %%A in ("%~1") do (
set "ln=%%A"
call :repl
)
)
move /y "%~1.new" "%~1" >nul
exit /b
:repl
set "ln=%ln:"=""%"
set "ln=%ln:^=^^%"
set "ln=%ln:&=^&%"
set "ln=%ln:|=^|%"
set "ln=%ln:<=^<%"
set "ln=%ln:>=^>%"
set "ln=%ln:,=^,^,%"
set ln=%ln:""="%
set "ln=%ln:"=""%"
set "ln=%ln:,,=;%"
set "ln=%ln:^,^,=,%"
set "ln=%ln:""="%"
setlocal enableDelayedExpansion
echo(!ln!
exit /b
The script should be able to process almost any valid CSV file input. The only restrictions are:
- Empty lines are stripped from the output (should not be a problem with CSV)
- Line lengths are limited to around 8 kb. The exact limit is dependent on how many intermediate substitutions must be performed.