Question

Possible Duplicate:
When to store images in a database(mySQL) and when not? (Linking image to item in database)
Which is the best method to store files on the server (in database or storing the location alone)?

What should I consider to best choose between storing MP3s/pdfs with a database.
Is it better to store the file as a BLOB in the DB? or is it better to just have a field that allows me to look at the filesystem?

The files in questions are not likely to exceed 15MiB, but all likely to be at least 1MiB, in case that’s a consideration.

The data will be accessed via PHP, if that makes a difference…

Edit: I’ll go with what I was already doing 🙂 resources on the filesystem with filenames in the database. Thanks or the input!

Answer 1

I’ve been working on a project for the last 3 years where large attachments were stored in the database.

Advantages

• Managing records in a database is easier and less code than having to handle associated files.
• Even with CakePHP’s behaviors to automate the task. The database is easier.
• Some projects have security requirements to keep data in the database (even if it doesn’t really improve security).
• Automated build processes can easily backup the data, because it’s already in the database anyway.

Disadvantages

• Taking snapshots of the database for development is a pain.
• After you pass several GBs in size. No matter what people tell you. You’ll start to sweat wondering if the DB engine can handle it.
• Dumping the database starts to take very long.
• Data migration to a different schema starts to require non-SQL solutions.

Myths About DB Blobs

• That your database can’t handle a large database file size.
• That one bad sector on the hard drive means you could lose all your blob data.
• That there is a performance cost to storing in the database.
• That it’s more secure.
• That reading from the DB and sending to the client is slower than reading a file and sending it.

All the above (from my experience) is false.

Facts About Large Blobs

• There will come a time when you no longer want it in the database.
• GUID record identifiers are your friend. Don’t use the record (int)ID in the file name.
• You don’t need to reverse find a record from a file name, but you need to find a file from a record. So you can just use the guid as the file name.
• It’s easier to backup/copy a file system and a small database, then just a large database.
• Don’t store the files in your webroot.