Perl/SQL query - grouping 'one to many' records

pcdj_d has asked for the wisdom of the Perl Monks concerning the following question:

I have a situation where I have a table of incidents and a table of actions. The relationship is one incident has many actions.

I require the need to display via a web page, a list of incidents IDs and their relevant actions thus:

Indicent 1
- Action 1
- Action 2
- Action 3
Indicent 2
- Action 4
Incident 3
- Action 5
- Action 6

Currrently, I just retrieve a set of actions and can sort them by incident ID. However I would like each action grouped by incident, and the incident ID to be displayed just the once before each group of actions.

I currently have in one subroutine:

my $sql =
"
SELECT
incident.id AS i_id, incident.severity, incident.summary,
action.id AS a_id, action.incident_id, action.description, action.note
+s
FROM
incident,
action
WHERE
incident.id = action.incident_id
AND action.due_date > $date (a specified date)
AND action.due_date < current_date()
ORDER BY sort_date ASC
"

$sth->prepare($sql);
$sth->execute();

while (my $href =$sth->fetchrow_hashref()) {
my $details={};
$details->{'Incident ID'} = $$href{'i_id'};
$details->{'Action ID'} = $$href{'a_id'};
$details->{'Severity'} = $$href{'severity'};
$details->{'Summary'} = $$href{'summary'};
and so on...

push(my @results, $details);
}
return @results
[download]

And in another subroutine:

my @results = "first subroutine"

foreach my $result (@results) {
my $output = $result->{'Incident ID'}.", ";
$output .= $result->{'Action ID'}.", ";
$output .= $result->{'Severity'}.", ";
$output .= $result->{'Summary'}.", ";
and so on...
$counter++
}
return $output;
[download]

This will return something along the following lines:

1, 1, Severe, A severe incident
1, 2, Severe, A severe incident
1, 3, Severe, A severe incident
2, 4, Minor, A minor incident
2, 5, Minor, A minor incident

I would like something like:

1, Severe, A severe incident
-> 1 (action id, with further columns from the action table)
-> 2
-> 3
2, Minor, A minor incident
-> 4
-> 5

What's the easiest way to do this?

Comment on Perl/SQL query - grouping 'one to many' records Select or Download Code

Replies are listed 'Best First'.
Re: Perl/SQL query - grouping 'one to many' records by ahmad (Hermit) on Dec 09, 2008 at 23:26 UTC
Just use SQL group by function. Few examples from MySQL website	[reply]
Re^2: Perl/SQL query - grouping 'one to many' records by pcdj_d (Initiate) on Dec 09, 2008 at 23:36 UTC
I was under the impression you couldn't 'group by' text fields? Which all these fields are (the figures '1,2,3' etc were just an example - they are all numeric IDs prefixed with some text.)	[reply]
Re^3: Perl/SQL query - grouping 'one to many' records by jasonk (Parson) on Dec 10, 2008 at 02:04 UTC
Your impression is wrong, you can group by any field you want. www.jasonkohles.com We're not surrounded, we're in a target-rich environment!	[reply]
Re: Perl/SQL query - grouping 'one to many' records by duckyd (Hermit) on Dec 10, 2008 at 00:47 UTC
I think the problem is mostly due to how you're storing the results as you fetch them from the database. If you use a hash keyed on incident.id and push the results for each action in, then the data will already be in the format you want to print it in. Here's an example: #!/usr/local/bin/perl use strict; use warnings; my @query = ( { i_id => 1, summary => 's 1', a_id => 1, description => 'a 1', }, { i_id => 1, summary => 's 1', a_id => 2, description => 'a 2', }, { i_id => 1, summary => 's 1', a_id => 3, description => 'a 3', }, { i_id => 2, summary => 's 2', a_id => 4, description => 'a 4', }, { i_id => 2, summary => 's 2', a_id => 5, description => 'a 5', }, ); my %results; # analagous to your while ( my $href = $sth->fetchrow_hashref() ){ while( my $href = shift @query ){ # treat each entry in %results as a hash ref unless( $results{ $href->{ i_id } } ){ # save the summary information in the summary key $results{ $href->{ i_id } }->{ summary } = $href->{ summary }; } # push a hash ref of action information into { actions } push @{ $results{ $href->{ i_id } }->{ actions } }, { a_id => $href->{ a_id }, desc => $href->{ description }, }; } # now it's grouped by incident, and you can print it: foreach my $i_id ( keys %results ){ print "Incident: $i_id - $results{ $i_id }->{ summary }\n"; foreach my $action ( @{ $results{ $i_id }->{ actions } } ){ print " $action->{ a_id }: $action->{ desc }\n"; } } [download] note that the last foreach, where the printing happens, is pretty straightforward - loop over the keys (one for each incident), print the incident info, and then loop through the actions for each. The output looks like this: `Incident: 1 - s 1 1: a 1 2: a 2 3: a 3 Incident: 2 - s 2 4: a 4 5: a 5` [download]	[reply] [d/l] [select]
Re^2: Perl/SQL query - grouping 'one to many' records by pcdj_d (Initiate) on Dec 11, 2008 at 09:14 UTC
Thanks duckyd, that worked a treat! One further question though - my results come out in a seemingly random order and cannot be sorted in any way. I believe this is due to the nature of hashes, and you can't guarantee in which order they are returned. Is there any simple way of modifying this so that I can have some control over how the results are sorted when displayed on screen? Something as simple as sorting by $i_id would be good!	[reply]


Think about Loose Coupling
	PerlMonks