Reputation: 11
I have a posts table like so:
+-----+----------+------------+------------+
| id | topic_id | text | timestamp |
+-----+----------+------------+------------+
| 789 | 2 | foobar | 1396026357 |
| 790 | 2 | foobar | 1396026358 |
| 791 | 2 | foobar | 1396026359 |
| 792 | 3 | foobar | 1396026360 |
| 793 | 3 | foobar | 1396026361 |
+-----+----------+------------+------------+
How would I could about "grouping" the results by topic id, while pulling the most recent record (sorting by timestamp desc)?
I've come to the understanding that I might not want "group_by" but rather "distinct on". My postgres query looks like this:
select distinct on (topic_id) topic_id, id, text, timestamp
from posts
order by topic_id desc, timestamp desc;
This works great. However, I can't figure out if this is something I can do in DBIx::Class without having to write a custom ResultSource::View. I've tried various arrangements of group_by with selects and columns, and have tried distinct => 1. If/when a result is returned, it doesn't actually preserve the uniqueness.
Is there a way to write the query I am trying through a resultset search, or is there perhaps a better way to achieve the same result through a different type of query?
Upvotes: 1
Views: 562
Reputation: 1464
Check out the section in the DBIC Cookbook on grouping results.
I believe what you want is something along the lines of this though:
my $rs = $base_posts_rs->search(undef, {
columns => [ {topic_id=>"topic_id"}, {text=>"text"}, {timestamp=>"timestamp"} ],
group_by => ["topic_id"],
order_by => [ {-desc=>"topic_id"}, {-desc=>"timestamp"} ],
})
Edit: A quick and dirty way to get around strict SQL grouping would be something like this:
my $rs = $base_posts_rs->search(undef, {
columns => [
{ topic_id => \"MAX(topic_id)" },
{ text => \"MAX(text)" },
{ timestamp => \"MAX(timestamp)" },
],
group_by => ["topic_id"],
order_by => [ {-desc=>"topic_id"}, {-desc=>"timestamp"} ],
})
Of course, use the appropriate aggregate function for your need.
Upvotes: 1