This might be quite easy but have no idea how to.
I have a table that can have repeated values for a particular non-key column field. How do I write a SQL query using Query Builder or Eloquent that will fetch rows with distinct values for that column?
Note that I am not fetching that column only, it is in conjunction with other column values, so distinct()
might not really work. So that question basically can be how to specify the column that I want to be distinct in a query now that distinct()
accepts no parameters?
You should use groupby
. In Query Builder you can do it this way:
$users = DB::table('users')
->select('id','name', 'email')
->groupBy('name')
->get();
In Eloquent you can also query like this:
$users = User::select('name')->distinct()->get();
Note that I am not fetching that column only, it is in conjunction with other column values
$users = User::select('column1', 'column2', 'column3')->distinct()->get();
in eloquent you can use this
$users = User::select('name')->groupBy('name')->get()->toArray() ;
groupBy is actually fetching the distinct values, in fact the groupBy will categorize the same values, so that we can use aggregate functions on them. but in this scenario we have no aggregate functions, we are just selecting the value which will cause the result to have distinct values
in_array()
function, it never works. To fix it, I tried ->lists()
instead (Version 5.2). So, $users = User::select('name')->groupBy('name')->lists('name');
worked fine for php's in_array()
;
Though I am late to answer this, a better approach to get distinct records using Eloquent would be
$user_names = User::distinct()->get(['name']);
get()
method (and I've tested also to the first()
method), which is equivalent to using the select()
method as shown in a few answers here. Though somehow groupBy
still seems to score the highest. However this would be truly distinct if that is the only column I'm selecting.
$user_names = User::distinct()->count(['name']);
also works
**
Tested for Laravel 5.8
**
Since you wanna get all columns from the table, you can collect all of the data and then filter it using Collections function called Unique
// Get all users with unique name
User::all()->unique('name')
or
// Get all & latest users with unique name
User::latest()->get()->unique('name')
For more information you can check Laravel Collection Documentations
EDIT: You might have issue with perfomance, by using Unique() you'll get all data first from User table, and then Laravel will filter it. This way isn't good if you have lots of Users data. You can use query builder and call each fields that you wanna use, example:
User::select('username','email','name')->distinct('name')->get();
Grouping by will not work if the database rules don't allow any of the select fields to be outside of an aggregate function. Instead use the laravel collections.
$users = DB::table('users')
->select('id','name', 'email')
->get();
foreach($users->unique('name') as $user){
//....
}
Someone pointed out that this may not be great on performance for large collections. I would recommend adding a key to the collection. The method to use is called keyBy. This is the simple method.
$users = DB::table('users')
->select('id','name', 'email')
->get()
->keyBy('name');
The keyBy also allows you to add a call back function for more complex things...
$users = DB::table('users')
->select('id','name', 'email')
->get()
->keyBy(function($user){
return $user->name . '-' . $user->id;
});
If you have to iterate over large collections, adding a key to it solve the performance issue.
Note that groupBy
as used above won't work for postgres.
Using distinct
is probably a better option - e.g. $users = User::query()->distinct()->get();
If you use query
you can select all the columns as requested.
$users = User::select('column1', 'column2', 'column3')->distinct()->get();
retrieves all three coulmns for distinct rows in the table. You can add as many columns as you wish.
I found this method working quite well (for me) to produce a flat array of unique values:
$uniqueNames = User::select('name')->distinct()->pluck('name')->toArray();
If you ran ->toSql()
on this query builder, you will see it generates a query like this:
select distinct `name` from `users`
The ->pluck()
is handled by illuminate\collection lib (not via sql query).
I had the same issues when trying to populate a list of all the unique threads a user had with other users. This did the trick for me
Message::where('from_user', $user->id)
->select(['from_user', 'to_user'])
->selectRaw('MAX(created_at) AS last_date')
->groupBy(['from_user', 'to_user'])
->orderBy('last_date', 'DESC')
->get()
// Get unique value for table 'add_new_videos' column name 'project_id'
$project_id = DB::table('add_new_videos')->distinct()->get(['project_id']);
Here are 3 ways I have tested that will give same result:
User::distinct()->get(['name'])->pluck('name');
User::select('name')->distinct()->pluck('name')->all();
DB::table('users')->select('name')->groupBy('name')->get()->pluck('name')->all();
For those who like me doing same mistake. Here is the elaborated answer Tested in Laravel 5.7
A. Records in DB
UserFile::orderBy('created_at','desc')->get()->toArray();
Array
(
[0] => Array
(
[id] => 2073
[type] => 'DL'
[url] => 'https://i.picsum.photos/12/884/200/300.jpg'
[created_at] => 2020-08-05 17:16:48
[updated_at] => 2020-08-06 18:08:38
)
[1] => Array
(
[id] => 2074
[type] => 'PROFILE'
[url] => 'https://i.picsum.photos/13/884/200/300.jpg'
[created_at] => 2020-08-05 17:20:06
[updated_at] => 2020-08-06 18:08:38
)
[2] => Array
(
[id] => 2076
[type] => 'PROFILE'
[url] => 'https://i.picsum.photos/13/884/200/300.jpg'
[created_at] => 2020-08-05 17:22:01
[updated_at] => 2020-08-06 18:08:38
)
[3] => Array
(
[id] => 2086
[type] => 'PROFILE'
[url] => 'https://i.picsum.photos/13/884/200/300.jpg'
[created_at] => 2020-08-05 19:22:41
[updated_at] => 2020-08-06 18:08:38
)
)
B. Desired Grouped result
UserFile::select('type','url','updated_at)->distinct('type')->get()->toArray();
Array
(
[0] => Array
(
[type] => 'DL'
[url] => 'https://i.picsum.photos/12/884/200/300.jpg'
[updated_at] => 2020-08-06 18:08:38
)
[1] => Array
(
[type] => 'PROFILE'
[url] => 'https://i.picsum.photos/13/884/200/300.jpg'
[updated_at] => 2020-08-06 18:08:38
)
)
So Pass only those columns in "select()"
, values of which are same. For example: 'type','url'
. You can add more columns provided they have same value like 'updated_at'
.
If you try to pass "created_at"
or "id"
in "select()"
, then you will get the records same as A. Because they are different for each row in DB.
Success story sharing
where()
it breaks, how can i fix this?group by
is not the same asdistinct
.group by
will change the behavior of aggregate functions likecount()
inside the SQL query.distinct
will not change the behavior of such functions, so you get different results depending on which one you use.