Skip to main content

How to Remove a Specific Column from a CSV File in Batch Script

CSV files are often over-designed for a specific task, for example, a user export might contain 20 columns of data (phone number, address, manager, etc.), but your script only needs the "Username" and "Email." Column removal is the process of stripping away these unnecessary vertical segments to create a leaner, focused dataset that is faster to process and easier to read.

In this guide, we will demonstrate how to remove a specific column using the tokens and delims parameters.

The Strategy: Targeted Tokenization​

If your CSV has 4 columns: A,B,C,D and you want to remove Column 2 (B):

  1. Set your delims to a comma (,).
  2. Set your tokens to capture only columns 1, 3, and 4.
  3. Rewrite the file using only those captured variables.

Given a CSV file with the following structure:

ID,Name,Email,Department
1001,Jane Smith,jane@example.com,Marketing
1002,Bob Jones,bob@example.com,Engineering
1003,Alice Park,alice@example.com,Finance

This script removes Column 2 (Name), producing ID,Email,Department.

@echo off
setlocal disabledelayedexpansion

set "Source=CompleteList.csv"
set "Output=StrippedList.csv"

:: Verify source file exists
if not exist "%Source%" (
    echo [ERROR] Source file "%Source%" not found.
    pause
    exit /b 1
)

echo Removing Column 2 from "%Source%"...

:: tokens=1,3,4 captures columns 1, 3, and 4, skipping column 2
:: %%A = Column 1 (ID)
:: %%B = Column 3 (Email) - next sequential letter after %%A
:: %%C = Column 4 (Department)
(
    for /f "usebackq tokens=1,3,4 delims=," %%A in ("%Source%") do (
        set "col1=%%A"
        set "col2=%%B"
        set "col3=%%C"
        setlocal enabledelayedexpansion
        echo(!col1!,!col2!,!col3!
        endlocal
    )
) > "%Output%"

echo [SUCCESS] Column removed. New CSV saved to "%Output%".
pause
exit /b 0

Output:

ID,Email,Department
1001,jane@example.com,Marketing
1002,bob@example.com,Engineering
1003,alice@example.com,Finance
warning

The for /f command assigns token variables to sequential letters starting from your loop variable, not to letters matching the token numbers. With tokens=1,3,4 and loop variable %%A, column 1 is %%A, column 3 is %%B, and column 4 is %%C. This is a common source of confusion, the variable letters do not correspond to the column positions.

tip

The script uses the delayed expansion toggle pattern: each column value is set with delayed expansion disabled to preserve literal ! characters in the data. The values are then output with delayed expansion enabled to safely handle &, |, >, <, and other special characters that would break a direct echo %%A,%%C,%%D approach.

Removing columns by index in Batch is fragile if a field contains a comma (e.g., "Doe, John"). PowerShell's Import-Csv engine is aware of quoted fields and header names, making it the correct tool for real-world CSV data.

@echo off
setlocal

set "Source=CompleteList.csv"
set "Dest=StrippedList.csv"
set "RemoveColumn=Name"

:: Verify source file exists
if not exist "%Source%" (
    echo [ERROR] Source file "%Source%" not found.
    pause
    exit /b 1
)

echo Removing column "%RemoveColumn%" from "%Source%"...

:: Import-Csv parses the CSV with full RFC 4180 compliance
:: Select-Object -ExcludeProperty removes the named column
:: Export-Csv writes the result back as a valid CSV
powershell -NoProfile -Command ^
    "Import-Csv -Path '%Source%' | " ^
    "Select-Object * -ExcludeProperty '%RemoveColumn%' | " ^
    "Export-Csv -Path '%Dest%' -NoTypeInformation -Encoding UTF8"

if %errorlevel% equ 0 (
    echo [SUCCESS] Column "%RemoveColumn%" removed. Saved to "%Dest%".
) else (
    echo [ERROR] Column removal failed.
    pause
    exit /b 1
)
pause
exit /b 0

Output:

"ID","Email","Department"
"1001","jane@example.com","Marketing"
"1002","bob@example.com","Engineering"
"1003","alice@example.com","Finance"
info

The Import-Csv cmdlet handles all standard CSV complexities: quoted fields containing commas, escaped quotes within fields, and multi-line values. The -ExcludeProperty parameter removes the column by its header name, which is more reliable and self-documenting than using column index numbers. If you need to remove multiple columns, list them separated by commas: -ExcludeProperty 'Address','Phone','ManagerID'.

Why Remove Columns?​

  1. Privacy/GDPR: Stripping PII (Personally Identifiable Information) like home addresses from a dataset before sharing it with a third-party tool.
  2. Performance: If a script needs to process a 10,000-line file, reducing the file from 20 columns to 2 columns significantly reduces memory usage and processing time.
  3. Cross-Tool Compatibility: If Tool A outputs 5 columns but Tool B only accepts 3, you must remove the extra data to prevent invalid format errors.

Best Practices​

  1. Verify Source File: Always check that the input file exists before processing. A missing file will cause the for /f loop to silently produce empty output.
  2. Quoted Fields: If any cell contains the delimiter character within its value (like "New York, NY"), the Batch delims=, will split that cell into two separate columns, corrupting all subsequent column positions on that row. Use Method 2 for any CSV that may contain quoted fields.
  3. Header Row: The Batch method processes the header row identically to data rows, it will output the header with the same column removed. However, if the header contains different content than expected, verify the output to ensure labels still match their data columns. The PowerShell method removes columns by header name, so header integrity is maintained automatically.
  4. Token Limits: Batch for /f loops support tokens up to index 31. If you need to remove a column beyond position 31, you must use PowerShell.
  5. Variable Column Count: If different rows in your CSV have different numbers of columns, the Batch method will silently misalign the output. The PowerShell method handles variable column counts gracefully because it parses by header name.
  6. Wildcard Token for Remaining Columns: If you need to keep all columns after a certain position, use the * wildcard token (e.g., tokens=1,3*). The * captures the remainder of the line as a single unsplit string, which preserves commas within the remaining content but prevents you from selectively removing columns from within that remainder.
  7. Encoding: Include -Encoding UTF8 in the PowerShell Export-Csv command to prevent non-ASCII characters from being corrupted.

Conclusion​

Removing a specific column is a vital part of data pruning. It ensures that your scripts are only working with the information they truly need, improving performance and security. Whether you use the manual token-skipping method in Batch for simple, well-structured files or the advanced exclusion logic in PowerShell for real-world CSV data, the ability to vertically filter your data is a hallmark of professional automation.

Table of Contents