Modifying CSV file and Preserving Order - perl

Question

The question that follows is a made up simplified example of a more complex problem that I'm trying to solve. I would like to preserve the structure of the code, especially the use of the %hash to store the outcomes for each patient but I do not need to read the data file into memory (but I cannot find a way of reading my csv data file line by line from the end.)
My sample data is made up of events that occur to patients. A patient can be added to the study (Event=B) or he can die (Event=D) or exit the study(Event=F.) Death and Exit are the only two possible outcomes for each patient.
For each event I have the date of occurrence (in hours from given point in time), the unique ID number of each patient, the event and the Outcome (a field set to 0 for every patient.)
I'm trying to write a code that will change the input file by putting next to each addition of a new patient, what is his eventual outcome (death or exit.)
In order to do so, I read the file from the end, and whenever I encounter a death or exit of a patient, I populate a hash that matches patient ID with outcome. When I encounter an event telling me that a new patient has been added to the study, I then match his ID with those in the hash and change the value of "Outcome" from 0 to either D or F.
I have been able to write a code that reads the file from bottom and then creates a new modified file with the updated value for Outcome. The problem is that since I read the input file from bottom to top and print each line after reading it, the output file is in reversed order and I do not know how to change this. Also, ideally I don't want to create a new file bu I would like to simply modify the input one. However, I have failed with every attempt to do so.
Sample data:
Data,PatientNumber,Event,Outcome
25201027,562962838335407,B,0
25201028,562962838335408,B,0
25201100,562962838335407,D,0
25201128,562962838335408,F,0
My code:
#!/usr/bin/perl
use strict;
use warnings;
use diagnostics;
open (my $fh_input, "<", "mini_test2.csv")
or die "cannot open > mini_test2.csv: $!";
my #lines = <$fh_input>;
close $fh_input;
open (my $fh_output, ">>", "Revised_mini_test2.csv")
or die "cannot open > Revised_mini_test2.csv: $!";
my $length = scalar(#lines);
my %outcome;
my #input_variables;
for (my $i = 1; $i < #lines; $i++){
chomp($lines[$length-$i]);
#input_variables=split(/,/, $lines[$length - $i]);
if ($input_variables[2] eq "D" || $input_variables[2] eq "F"){
$outcome{$input_variables[1]} = $input_variables[2];
my $line = join(",", #input_variables);
print $fh_output $line . "\n";
}
elsif($input_variables[2] eq "B") {
$input_variables[3]=$outcome{$input_variables[1]};
my $line = join(",", #input_variables);
print $fh_output $line . "\n";
}
else{
# necessary since the actual data has many more possible "Events"
my $line = join(",", #input_variables);
print $fh_output $line . "\n";
}
}
close $fh_output;
EDIT: desired output should be
Data,PatientNumber,Event,Outcome
25201027,562962838335407,B,D
25201028,562962838335408,B,F
25201100,562962838335407,D,0
25201128,562962838335408,F,0
Also, an additional complication is that the unique patient ID after the exit of a patient gets re-used. This means that I cannot do a 1st pass and store the outcome for each patient and a 2nd one to update the values of Outcome.
EDIT 2: let me clarify that when I say that each patient has a "unique ID" I mean that there cannot be in the study, at the same time, two patients with the same ID. However, if a patient exits the study, his ID gets re-used.

Answer 1

Update
I have just read your additional information that patient numbers are re-used once they exit the study. Why you would design a system like that I don't know, but there it is
It becomes far harder to write something straightforward without reading the file into an array, so that's what I have done here
use strict;
use warnings;
use 5.010;
use autodie;
open my $fh, '<', 'mini_test2.csv';
my #data;
while ( <$fh> ) {
chomp;
push #data, [ split /,/ ];
}
my %outcome;
for ( my $i = $#data; $i > 0; --$i ) {
my ($patient_number, $event) = #{$data[$i]}[1,2];
if ( $event =~ /[DF]/ ) {
$outcome{$patient_number} = $event;
}
elsif ( $event =~ /[B]/ ) {
$data[$i][3] = delete $outcome{$patient_number} // 0;
}
}
print join(',', #$_), "\n" for #data;
output
Data,PatientNumber,Event,Outcome
25201027,562962838335407,B,D
25201028,562962838335408,B,F
25201100,562962838335407,D,0
25201128,562962838335408,F,0
There are a few ways to approach this. I have chosen to take two passes through the file, first accumulating the outcome for each patient in a hash, and then replacing all the outcome fields in the B records
use strict;
use warnings;
use 5.010;
use autodie;
use Fcntl ':seek';
my %outcome;
open my $fh, '<', 'mini_test2.csv';
<$fh>; # Drop header
while ( <$fh> ) {
chomp;
my #fields = split /,/;
my ($patient_number, $event) = #fields[1,2];
if ( $event =~ /[DF]/ ) {
$outcome{$patient_number} = $event;
}
}
seek $fh, 0, SEEK_SET; # Rewind
print scalar <$fh>; # Copy header
while ( <$fh> ) {
chomp;
my #fields = split /,/;
my ($patient_number, $event) = #fields[1,2];
if ( $event !~ /[DF]/ ) {
$fields[3] = $outcome{$patient_number} // 0;
}
print join(',', #fields), "\n";
}
output
Data,PatientNumber,Event,Outcome
25201027,562962838335407,B,D
25201028,562962838335408,B,F
25201100,562962838335407,D,0
25201128,562962838335408,F,0

Answer 2

What we can do is instead of printing out the line at each stage, we'll write it back to the array of lines. Then we can just print them out at the end.
for (my $i=$#lines; i>=0; i--)
{
chomp $lines[$i];
#input_variables = split /,/, $lines[$i];
if ($input_variables[2] eq "D" || $input_variables[2] eq "F")
{
$outcome{$input_variables[1]}=$input_variables[2];
}else
{
$input_variables[3]=$outcome{$input_variables[1]};
}
$line[$i] = join ",", #input_variables;
}
$, = "\n"; #Make list seperator for printing a newline.
print $fh_output #lines;
As for the second question of modifying the original file. It is possible to open a file for both reading and writing using modes "+<", "+>", or "+>>". Don't do this! It is error prone as you must replace data character by character.
The standard way to "modify" an existing file is to rename it, read from the renamed file, write to a new file with the original name, and delete the temp file.
my $file_name = "mini_test2.csv";
my $tmp_file_name = $file_name . ".tmp";
rename $file_name, $tmp_file_name;
open (my $fh_input, "<", $tmp_file_name)
or die "cannot open > $tmp_file_name: $!";
open (my $fh_output, ">>", $file_name)
or die "cannot open > $file_name: $!";
#Your code to process the data.
close $fh_input;
close $fh_output;
#delete the temp file
unlink $tmp_file_name;
But, in your case, you slurp all of the data into memory right away. Just open for writing that clobbers existing files
open (my $fh_output, ">", "mini_test2.csv")
or die "cannot open > mini_test2.csv: $!";