Between the Manticore query and the final result to the user there can be additional processing. As in most cases, the interest is to sort by a relevance score, it’s important to not lose this sorting.

A typical workflow using Manticore Search along with a database looks like this:

• perform Manticore query
• from result set get the list of document ids
• perform a database query using the document ids to get final result for output

Getting the corresponding records from the database can be made in a single query using the IN(ids) predicate.

mysql> SELECT * FROM mytable WHERE id IN(20,10,3,4,6)

This type of query is very fast, but the result set may not be in the order we would expect. Since there is no ordering set, the database engine will simply add the records to the result set as soon as it finds them while searching them in the table. This means we need to add explicit ordering to the query to preserve the same order we received from the Manticore query.

In MySQL we can use FIELD() function to sort the set the way we want:

mysql> SELECT * FROM mytable WHERE id IN(20,10,3,4,6) ORDER BY FIELD(id,20,10,3,4,6).

Similar methods exists in other databases. In PostgreSQL it can be done in several ways, like using ORDER BY CASE, WITH ORDINALITY (9.4+) or array_position() (9.6+).

As alternative, keeping the order from the Manticore result can also be made in the application code. A simplified code can look like this:

$ids = [];
foreach($ms_conn($manticore_query as $row)
{
   $ids[] = $row['id'];
}
$tmpdocs = [];
foreach($db_conn->query($database_query) as $row)
{
   $tmpdocs[$row['id']] = $row
}
$final_result = [];
foreach($ids as $id)
{
   $final_result[] = $tmpdocs[$id];
}

As conclusion, if you retrieve data from a repository using ids from a Manticore query make sure to not lose the order provided by the Manticore query result set.