Powershell / Perl : Merging multiple CSV files into one?

Question

I have the following CSV files, I want to merge these into a single CSV

01.csv

apples,48,12,7
pear,17,16,2
orange,22,6,1

02.csv

apples,51,8,6
grape,87,42,12
pear,22,3,7

03.csv

apples,11,12,13
grape,81,5,8
pear,11,5,6

04.csv

apples,14,12,8
orange,5,7,9

Desired output:

apples,48,12,7,51,8,6,11,12,13,14,12,8
grape,,,87,42,12,81,5,8,,,
pear,17,16,2,22,3,7,11,5,6,,,
orange,22,6,1,,,,,,5,7,9

Can anyone provide guidance on how to achieve this? Preferably using Powershell but open to alternatives like Perl if that's easier.

Thanks Pantik, your code's output is close to what I want:

apples,48,12,7,51,8,6,11,12,13,14,12,8
grape,87,42,12,81,5,8
orange,22,6,1,5,7,9
pear,17,16,2,22,3,7,11,5,6

Unfortunately I need "placeholder" commas in place for when the entry is NOT present in a CSV file, e.g. orange,22,6,1,,,,,,5,7,9 rather than orange,22,6,1,5,7,9

UPDATE: I would like these parsed in order of the filenames, e.g.:

$myFiles = @(gci *.csv) | sort Name
foreach ($file in $myFiles){

regards ted

It looks like you want the data ordered by filenames. For example, you have empty records in `orange` from `2.csv` and `3.csv`. If this is a requirement, you should add that to the question. — TLP, May 13 '11 at 08:55

score 4 · Answer 1 · answered May 13 '11 at 09:32

Here is my Perl version:

use strict;
use warnings;

my $filenum = 0;

my ( %fruits, %data );
foreach my $file ( sort glob("*.csv") ) {

    $filenum++;
    open my $fh, "<", $file or die $!;

    while ( my $line = <$fh> ) {

        chomp $line;

        my ( $fruit, @values ) = split /,/, $line;

        $fruits{$fruit} = 1;

        $data{$filenum}{$fruit} = \@values;
    }

    close $fh;
}
foreach my $fruit ( sort keys %fruits ) {

    print $fruit, ",", join( ",", map { $data{$_}{$fruit} ? @{ $data{$_}{$fruit} } : ",," } 1 .. $filenum ), "\n";
}

Which gives me:

apples,48,12,7,51,8,6,11,12,13,14,12,8
grape,,,,87,42,12,81,5,8,,,
orange,22,6,1,,,,,,,5,7,9
pear,17,16,2,22,3,7,11,5,6,,,

So do you have a typo for grape or i have misunderstood something?

Your default sort will not do for filenames beyond `9.csv` though, as `11.csv` will come before `2.csv`. — TLP, May 13 '11 at 09:45
Thanks gangablass... yes in my desired output I have a typo , there should not be a space immediately after grape. Have updated. — ted, May 13 '11 at 10:05

score 2 · Answer 2 · answered May 13 '11 at 10:37

Ok, gangabass solution works, and is cooler than mine, but I'll add mine anyway. It is slightly stricter, and preserves a data structure that can be used as well. So, enjoy. ;)

use strict;
use warnings;

opendir my $dir, '.' or die $!;
my @csv = grep (/^\d+\.csv$/i, readdir $dir);
closedir $dir;
# sorting numerically based on leading digits in filename
@csv = sort {($a=~/^(\d+)/)[0] <=> ($b=~/^(\d+)/)[0]} @csv;

my %data;

# To print empty records we first need to know all the names
for my $file (@csv) {
    open my $fh, '<', $file or die $!;
    while (<$fh>) {
        if (m/^([^,]+),/) {
            @{ $data{$1} } = ();
        }
    }
    close $fh;
}

# Now we can fill in values
for my $file (@csv) {
    open my $fh, '<', $file or die $!;
    my %tmp;
    while (<$fh>) {
        chomp;
        next if (/^\s*$/);
        my ($tag,@values) = split (/,/);
        $tmp{$tag} = \@values;
    }
    for my $key (keys %data) {
        unless (defined $tmp{$key}) {
            # Fill in empty values
            @{$tmp{$key}} = ("","","");
        }
        push @{ $data{$key} }, @{ $tmp{$key} };
    }
}

&myreport; 

sub myreport {
    for my $key (sort keys %data) {
        print "$key," . (join ',', @{$data{$key}}), "\n";
    }
}

mjolinor · Answer 3 · 2011-05-13T13:29:51.197

2

Powershell:

$produce = "apples","grape","orange","pear"
$produce_hash = @{}
$produce | foreach-object {$produce_hash[$_] = @(,$_)}

$myFiles = @(gci *.csv) | sort Name
 foreach ($file in $myFiles){ 
    $file_hash = @{}
    $produce | foreach-object {$file_hash[$_] = @($null,$null,$null)}
        get-content $file | foreach-object{
            $line = $_.split(",")
            $file_hash[$line[0]] = $line[1..3]
            }
    $produce | foreach-object {
        $produce_hash[$_] += $file_hash[$_]
        }
  }

$ofs = ","
$out = @()
$produce | foreach-object {
 $out += [string]$produce_hash[$_]
 }

$out | out-file "outputfile.csv" 

gc outputfile.csv

apples,48,12,7,51,8,6,11,12,13,14,12,8
grape,,,,87,42,12,81,5,8,,,
orange,22,6,1,,,,,,,5,7,9
pear,17,16,2,22,3,7,11,5,6,,,

Should be easy to modify for additional items. Just add them to the $produce array.

edited May 13 '11 at 13:29

answered May 13 '11 at 11:17

mjolinor

66,130
7
114
135

thanks mjolinor, can this modified so that you don't need to manually enter the items in the $produce array... as it may not be known in advance what the items will be... – ted May 13 '11 at 13:45
Possible. 2 ways I see to do that: 1- read through the data twice, using the first pass to collect the unique values of the first element to build the $produce array. 2 - set up a counter and increment as you process each file so you know you may arrays of $nulls you need to add before the first set of values you got for that item. Which one works best may depend on how many / how big your data files are. – mjolinor May 13 '11 at 13:54
Posted a second solution that auto-populates $produce – mjolinor May 13 '11 at 15:05

score 2 · Answer 4 · answered May 13 '11 at 15:04

Second Powershell solution (as requested)

   $produce = @()
   $produce_hash = @{}
    $file_count = -1
    $myFiles = @(gci 0*.csv) | sort Name
     foreach ($file in $myFiles){ 
        $file_count ++
        $file_hash = @{}
                get-content $file | foreach-object{
                $line = $_.split(",")

                if ($produce -contains $line[0]){
                    $file_hash[$line[0]] += $line[1..3]
                    }

                else {
                    $produce += $line[0]
                    $file_hash[$line[0]] = @(,$line[0]) + (@($null) * 3 *  $file_count) + $line[1..3]
                    }

                  }
              $produce | foreach-object { 
                if ($file_hash[$_]){$produce_hash[$_] += $file_hash[$_]} 
                else {$produce_hash[$_] += @(,$null) * 3}
               }

    }          

    $ofs = ","
    $out = @()
    $produce_hash.keys | foreach-object {
     $out += [string]$produce_hash[$_]
     }

    $out | out-file "outputfile.csv" 

    gc outputfile.csv
apples,48,12,7,51,8,6,11,12,13,14,12,8
grape,,,,87,42,12,81,5,8,,,
orange,22,6,1,,,,,,,5,7,9
pear,17,16,2,22,3,7,11,5,6,,,

Tomas Panik · Answer 5 · 2011-05-13T18:43:41.260

you have to parse the files, I don't see easier way hot to do it

solution in powershell:

UPDATE: ok, adjusted a bit - hopefully understandable

$items = @{}
$colCount = 0 # total amount of columns
# loop through all files
foreach ($file in (gci *.csv | sort Name))
{
    $content = Get-Content $file
    $itemsToAdd = 0; # columns added by this file
    foreach ($line in $content)
    {
        if ($line -match "^(?<group>\w+),(?<value>.*)") 
        { 
            $group = $matches["group"]
            if (-not $items.ContainsKey($group)) 
            {   # in case the row doesn't exists add and fill with empty columns
                $items.Add($group, @()) 
                for($i = 0; $i -lt $colCount; $i++) { $items[$group] += "" }
            }

            # add new values to correct row
            $matches["value"].Split(",") | foreach { $items[$group] += $_ }
            $itemsToAdd = ($matches["value"].Split(",") | measure).Count # saves col count
        } 
    }

    # in case that file didn't contain some row, add empty cols for those rows
    $colCount += $itemsToAdd
    $toAddEmpty = @()
    $items.Keys | ? { (($items[$_] | measure).Count -lt $colCount) } | foreach { $toAddEmpty += $_ }
    foreach ($key in $toAddEmpty) 
    {   
        for($i = 0; $i -lt $itemsToAdd; $i++) { $items[$key] += "" }
    }
}

# output
Remove-Item "output.csv" -ea 0
foreach ($key in $items.Keys)
{
    "$key,{0}" -f [string]::Join(",", $items[$key]) | Add-Content "output.csv"
}

Output:

apples,48,12,7,51,8,6,11,12,13,14,12,8
grape,,,,87,42,12,81,5,8,,,
orange,22,6,1,,,,,,,5,7,9
pear,17,16,2,22,3,7,11,5,6,,,

Thanks PantikT for this effort, much appreciated - please see my update to the my question for feedback, as this does not quite generate the output I'm looking for. — ted, May 13 '11 at 09:07

score 0 · Answer 6 · answered May 13 '11 at 12:38

Here is a more consise way how to do it. However, it still doesn't add the commas when the item is missing.

Get-ChildItem D:\temp\a\ *.csv | 
    Get-Content |
    ForEach-Object -begin { $result=@{} } -process {
        $name, $otherCols = $_ -split '(?<=\w+),'
        if (!$result[$name]) { $result[$name] = @() }
        $result[$name] += $otherCols
    } -end {
        $result.GetEnumerator() | % {
            "{0},{1}" -f $_.Key, ($_.Value -join ",")
        }
    } | Sort

Powershell / Perl : Merging multiple CSV files into one?

6 Answers6

Linked