I want to built a generic code to remove columns from file. Column number will be passed at run time.I may have multiple number like 2,5,10 so Can you please help.
I know how to remove when we know column number but I want a generic code where column number will be passed at run time.
I just use cut command or awk to print the columns as mentioned by sam05121988 but I just thought of making it generic so that it should work for any file just we need to pass column number which we need to drop. so I do not have idea on how to make it generic thats why I have posted here as experts are here.
Let's be clear about how you answered my questions...
I asked what operating system and shell you're using. If someone asked me that question, I would say Apple OS X El Capitan Version 10.11 and ksh version 93u+ 2012-08-01 . Your answer: "I am using Unix as OS." tells us that you may be using one of hundreds of releases of AIX, HP/UX, OS X, Solaris, or of several older operating systems and no indication at all of what shell nor version of the shell that you're using. (But, at least we do know that you are not using a Linux distribution and you are not using a BSD-based system.)
And, I asked what have you tried to solve this problem? Where are you stuck? Your answer: "I just thought of making it generic so that it should work for any file just we need to pass column number which we need to drop. so I do not have idea on how to make it generic thats why I have posted here as experts are here." I interpret that to mean you haven't tried anything and you want us to do the work for you.
You said you used awk as suggested by sam05121988, but sam05121988 suggested using cut ; not awk . Do you have awk code that does what you want to do for a non-generic list of fields to be removed from a non-generic file? If so show us that and we can help you modify it to work with any CSV file named at run-time to remove a list of fields specified at run time. Please just show us that you have made some attempt to produce some of the output you say you want from the sample input you have shown us.
I am not able to make it generic because I am not sure about total number of column in file as it may vary so Just wanted to know is there any way or option in above command to mention specific column number in command to drop instead of giving column number to be printed.
Thanks Don , Solution is working fine but Need clarification for below
IAm=${0##*/} - I have understand that it is something like $0 but here we are removing leading string if it matches with pattern */
printf '%s\n' "$@" - It is converting horizontal argument to verital(like Column to Rows) but It would be helpful if you can explain how awk works after passing these values to awk.
I have tried to understand with manual in filesystem but could not understand.
The script sets IAm to the basename of the name of the script. The shell expands $0 to the pathname used to invoke your script. Let's assume your script is named removecols and is located in your $HOME/bin directory. If you invoke your command with the command line:
$HOME/bin/removecols myfile.csv 4 2
the script sets IAm by throwing away everything from the beginning of the string through the last / character leaving just removecols as the string to be substituted in the usage message. And, if $HOME/bin is in your PATH variable and you invoke your script using the command:
removecols file1 2
The shell sets $0 to removecols when it starts and IAm gets the same string since there are no / characters in the expansion of $0 .
The shift command throws away the 1st positional parameter after file=$1 saves the name of the file to be processed that you passed to your script as the 1st positional parameter. So, after the shift , all of the remaining positional parameters are the numbers of the fields you want to delete. And "$@" expands to a list of each positional parameter as a quoted string with each parameter treated as a separate argument. So, for the 1st command-line shown above, the shell executes the command: printf '%s\n' "4" "2"
which produces the output:
4
2
which is fed as standard input through the pipeline into the awk script. And the - as the 2nd operand passed to the awk script tells it to read its standard input as its 1st input file.
The awk code:
BEGIN { FS = OFS = ","
}
tells awk to set the input field separator ( FS ) and the output field separator ( OFS ) to a comma before reading any lines from any input files.
The awk code:
FNR == NR {
d[$1]
next
}
says that for each line where the line # in the current input file ( FNR ) is the same as the line # of all lines read ( NR ) (which is true only for lines read from the 1st input file) create an array element with the value in the 1st field on that line as a subscript ( d[$1] ) and the awk next command tells the script to read the next input line and start again.
For lines read from the second input file (where FNR is not the same value as NR ) we execute the remaining commands in the awk script:
which sets ofs to an empty string (note that awk variable names are case sensitive). Then for each field in the current input line starting with the first field ( i = 1 ) and continuing while i is less than the number of fields in that line ( NF ), and incrementing i by one ( i++ ) each time through the loop, it checks to see if the field number is a subscript in the array containing being used as the list of fields to delete ( (i in d) ) and if it is not ( ! ), then it prints the current value of ofs followed by the current field ( $i ) and then sets ofs to the current output field separator ( OFS ). (So the first field printed will not have a field separator printed before it and all fields printed after that will print a comma as a field separator. After the loop completes the awk command: