perl -e'@cols =qw (need1 dont1 need2 dont2 need3); for (@cols) { push(
+@req,$_) unless /dont.*/;} print join (",",@req),$/;'
__END__
need1,need2,need3
[download]

You mentiioned regex so the above should help you. However if the names you want to choose are sent as input then you might want to construct a hash from the list and then drop the ones that should be excluded by checking the hash

.

The problem is that I don't nessessarily know what all the column names are. I know that any column that ends with "Bandwidth" or has "MSTCPLoopback" in it I don't want.

I'm trying to approach this from:


shift @cols; #Remove the first column
for (@cols) {
   delete $cols[$regex here]
   }
[download]

Then start my sql stuff.

You can just grep out the ones you want.

my @cols = ...;
shift @cols;  # blindy throw away first column
@cols = grep( $_ !~ /(Bandwidth|MSTCPLoopback)$/ , @cols);  # exclude 
+ones ending with "Bandwidth" or "MSTCPLoopback"
[download]

re: your code, note that delete is for hashes .. if you wanted to modify your code, you would use the slice function, but grep is much more powerful and perl-ish.

I am not going to do Re:Re:Re now as it makes it hard to read. This reply is for this node Re^8: Pre-process csv files before using

Could you please wrap your column header info inside code tags? Long lines don't wrap otherwise. Thanks!

#!/usr/bin/perl
use strict;
use warnings;

my $str = <DATA>;
my @origcols = split /,/,$str;
my @cols = ();

foreach (@origcols) {
    $_ =~ s/\.//g;
    push ( @cols , $_ ) unless /Bandwidth.*|MSTCPLoop.*/ ;
}

map {print $_,$/ } @cols;

__DATA__
PDHCSV40EasternDaylightTime.240,ERWW.COMMUNITIES.MemoryPagesPER.sec,ER
+WWCOMMUNITIESNetwor
kInterfaceEthernetAdapterModuleBytesTotalPERsec,ERWWCOMMUNITIESNetwork
+InterfaceEthernetAd
apterCurrentBandwidth
[download]

Output

PDHCSV40EasternDaylightTime240
ERWWCOMMUNITIESMemoryPagesPERsec
ERWWCOMMUNITIESNetworkInterfaceEthernetAdapterModuleBytesTotalPERsec
[download]

as you could see the periods are gone and no warnings. Also the column with bandwidth is not listd. Can you make sure your origcols array is correctly populated?

-SK

The code I gave you above should do what you are looking for.

perl -e'@cols =qw (field1 fieldband field3 fieldMSTCP okfield5); for (
+@cols) { push(@req,$_) unless /band.*|MSTCP.*/;} print join (",",@req
+),$/;'
__END__
field1,field3,okfield5
[download]

Here i am excluding names with band and MSTCP field names. Since I don't have any field names I am just hard coding it here

. -SK