NEW Create DBClassNameVarchar

[emteknetnz]

Issue #11358

[GuySartorelli]

@emteknetnz I saw you doing some performance tests of enum vs varchar, but I don't see the results of that anyway. Can you please dump those in a comment somewhere so I can validate your results?

[emteknetnz]

I used the code in #11358 (comment) slightly modified to generate 1,000,000 rows and to set a fairly standard looking FQCN

// ...
        $num = 1000000; // total number of records that should exist (note cannot decrease)
        $insert = 100000; // max inserts per query
        $count = self::get()->count();
        $diff = $num - $count;
        $t = '2024-08-29 17:47:10';
        parent::requireDefaultRecords();
        if ($diff > 0) {
            # use raw SQL to make insertions much faster
            $loops = ceil($diff / $insert);
            for ($loop = 0; $loop < $loops; $loop++) {
                $sqlBase = [];
                $sqlA = [];
                for ($i = 1; $i <= $insert && $i <= $diff; $i++) {
                    $id = $count + ($loop * $insert) + $i;
                    $sqlBase[] = "($id, 'App\\\\Src\\\\Something\\\\MyDataObjectA', 'My Data Object $id', '$t', '$t')";
                    $sqlA[] = "($id, 'Some value $id')";
                }
                if (!empty($sqlBase)) {
                    DB::query('INSERT INTO "MyBaseDataObject" ("ID", "ClassName", "Title", "LastEdited", "Created") VALUES ' . implode(',', $sqlBase) . ';');
                    DB::query('INSERT INTO "MyDataObjectA" ("ID", "SomeField") VALUES ' . implode(',', $sqlA) . ';
                    ');
                }
                $diff -= $insert;
            }
        }
    }
// ...

I then used the config to switch between enum and varchar, and then ran the following SQL to get the database size - note you may need to drop the database before switching as the ClassName isn't a valid class

SELECT table_schema "DB Name", ROUND(SUM(data_length + index_length) / 1024 / 1024, 1) "DB Size in MB" FROM information_schema.tables GROUP BY table_schema;

enum
+--------------------+---------------+
| DB Name            | DB Size in MB |
+--------------------+---------------+
| information_schema |           0.2 |
| mysql              |          10.3 |
| performance_schema |           0.0 |
| SS_mysite          |         118.6 |
+--------------------+---------------+

varchar
+--------------------+---------------+
| DB Name            | DB Size in MB |
+--------------------+---------------+
| information_schema |           0.2 |
| mysql              |          10.3 |
| performance_schema |           0.0 |
| SS_mysite          |         180.8 |
+--------------------+---------------+
4 rows in set (0.01 sec)

At first glance it's a huge relative % increase, though ignore that because there's almost no other data in the database so the ClassName column is a large proportion of it. In the real world this won't be the case

Size difference is 180.8 - 118.6 = 62.2mb per 1,000,000, or 6.2 mb per 100,000.

I realised the first time I did that I didn't have quad slashes in my script so the slashes weren't in the database so I had slightly less data stored. I'll bump up the estimate to 7mb per 100,000 in the docs to be conservative

[GuySartorelli]

After I've changed to varchar, if I add or remove a DataObject subclass, I still see the following in dev/build which are clearly alter table queries:

Field ChangeSetItem.ObjectClass: changed to enum( ........ ) character set utf8mb4 collate utf8mb4_unicode_ci default 'SilverStripe\\Assets\\File' (from enum( ....... )

Field FileLink.ParentClass: changed to enum( ........ ) character set utf8mb4 collate utf8mb4_unicode_ci default 'SilverStripe\\Assets\\File' (from enum( ....... )

Field SiteTreeLink.ParentClass: changed to enum( ........ ) character set utf8mb4 collate utf8mb4_unicode_ci default 'SilverStripe\\Assets\\File' (from enum( ....... )

The first one is particularly concerning, because change set items are created every time anything is published, which means there will be a lot of those records.
We should probably either account for that somehow in the code, or else add examples of how to update these as well in the docs.

[emteknetnz]

Need some additional config to get DBPolymorphicForeignKey to use varchar, have updated code sample + docs

[GuySartorelli]

LGTM