Super User is a question and answer site for computer enthusiasts and power users. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I have nine columns in my current .csv file, however, the 2nd and 3rd columns are blank and I want to get rid of them. Is there a way to use batch commands to remove these empty columns from my .csv or export a new .csv without those blank columns?

share|improve this question
Please edit your question to confirm: (1) Are you running Windows? (2) If so, are you looking for a pure Windows batch solution, or are you open to other alternatives? – Gord Thompson Mar 17 '13 at 14:18
up vote 1 down vote accepted

For /F is capable of parsing a .csv file. In one line, to be run directly from the command prompt:

for /F "tokens=1,2,3* delims=," %i in (input.csv) do @echo %i,%l>>output.csv

This traverses input.csv line for line and using a comma as delimiter, stores the first value in %i, the second in %j, the third in %k and the rest of the string (by using an asterisk in the list of tokens) in %l. By printing all but the second and third values, you are effectively taking those out.

However, for /F does not count empty strings as tokens, so the command cannot differentiate between two subsequent commas and a single one. This solution does work if the empty columns have a whitespace character, but may skip the wrong columns otherwise.

To circumvent this restriction, we can write a batch file that uses string replacement to insert a space between two subsequent commas. This replacement is executed twice to cope with the possibility of three commas. The quotes in ^"!line!^" are required to indicate for /F has to operate on a string instead of a file and are escaped to prevent !line! from being evaluated as a literal string instead of a variable.

@echo off
setlocal EnableDelayedExpansion
for /F "delims=" %%a in (input.csv) do (
    set line=%%a
    set line=!line:,,=, ,!
    set line=!line:,,=, ,!
    for /F "tokens=1,2,3* delims=," %%i in (^"!line!^") do (
        echo %%i,%%l>>output.csv

The result is a new .csv file, identical to input.csv, except for the omitted second and third columns.

share|improve this answer
Works like a charm. Thanks Marcks! – batchnoober Mar 18 '13 at 22:09
Note that this will add blanks to empty fields in columns other than 2 and 3. For example, The,quick,brown,fox,,,the,lazy,dog will become The,,,fox, , ,the,lazy,dog. – Scott Apr 11 '13 at 17:10

You did not specify which OS. If yours comes with cut, use this.

cut -d',' -f1,4-9

-d sets the delimiter, in this case to a ,. Since comma is not a special character, you can omit the quotes.
-f selects the fields, in this case field 1 and fields 4 till 9 (skipping fields 2 and 3).

Testfile "example.cvs"
foo1,   ,   , ba1, baz1, blah1, wibble1, wubble1
foo2,   ,   , ba2, baz2, blah2, wibble2, wubble2
foo2,   ,   , ba3,     , blah3, wibble3, wubble3
foo2,   ,   , ba4, baz4, blah4, wibble4, wubble4

An actual test of the command:

> cut -d',' -f1,4-9 example.cvs

foo1, ba1, baz1, blah1, wibble1, wubble1
foo2, ba2, baz2, blah2, wibble2, wubble2
foo2, ba3,     , blah3, wibble3, wubble3
foo2, ba4, baz4, blah4, wibble4, wubble4

Cut comes preinstalled with most operating systems (e.g., OS/X, Linux distributions, BSDs, Solaris, ...; just about everyone except DOS and Windows).

If you have Windows, you can:

  1. Open the CVS file in a spreadsheet (e.g., OpenOffice Calc or MS Excel), delete two columns, and then save the file. Hard to automate though, so use only if you need to do it once or rarely.
  2. or download Unix tools for Windows and put the cut command in a batch file.
share|improve this answer
My OS is Windows. Unfortunately, I would want this as an automated process since this is a .csv that's created everyday in the same format. Are there other alternatives for windows OS? – batchnoober Mar 18 '13 at 15:07
You can install a cut command (e.g. from ), or you can try something with for. An example of using for is for /f "tokens=5 delims= " %i in (file.txt) DO echo %i – Hennes Mar 18 '13 at 15:24
Unless you know the maximum number of columns, use -f1,4- to include all columns after the third. 4- means 4-∞. – Scott Apr 11 '13 at 18:11
True. I based my -f4-9 on the first sentence of the OP post "I have nine columns", but -f4- is more flexible less less typing. An other option would have been to used sed. – Hennes Apr 11 '13 at 18:22

I would open the CSV file in Notepad (or any text editor) and search and replace ,, with ,. This, of course, is assuming the data in your columns do not have ,, in them.

share|improve this answer
That will corrupt the structure of the file if any of the other columns have any empty values. – Gord Thompson Mar 17 '13 at 14:10

Assuming you're on Windows, the file has no headers, and you have Powershell available to you, you can use something like the following:

Import-Csv .\InFile.csv -Header col1,col2,col3,col4,col5 -Delimiter ',' | Select-Object col1,col4,col5 | ConvertTo-Csv | Select -skip 2 > .\OutFile.csv

So it:

  • Imports the CSV.
  • Assigns column names for reference.
  • Selects all 5 columns, except 2 and 3.
  • Converts back to a proper CSV.
  • Selects that CSV, starting at line 2 (to skip the assigned column headers)
  • Pipes it out to an output file.
share|improve this answer

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .