What's the best way to go about inserting data and images in bulk into MySQL database tables with PHP or MySQL?

While CSV is faster than XML, it is more error prone as there is no real way to validate it without writing code to count number of delimiters on each row. With XML, the script simply won't open an invalid file. I wrote this guide for my last employer on how to integrate an XML feed into a MySQL database using PHP. Hope it helps you in some way. http://files.perfiliate.com/solutions/BuyAtParseProductFeedsGuide.pdf

Depend on the amount of data you will insert. As Saurabh said, LOAD DATA INFILE is the fastest way. Turning off indexes and rebuilding them at the end could be faster than updating them for each record you insert. If is really heavy the amount of data, assuming you want to optimize inserts over queries, would be having the tables using the tokudb storage engine.

The safest way would be to build a parser (in PHP) to validate the data, then insert as per Bryon's answers. The quickest way is via MySQL's shell using the "load data" command to append to your table. http://dev.mysql.com/doc/refman/5.1/en/load-data.html But: I recommend not inserting images into your db. It's faster to serve static files from the file-system and easier to backup small text-only databases. If you choose to insert the images into the database, watch out for problems with large images due to MySQL/PHP message-buffer-size settings.

Look into using a migration tool like <a href="https://github.com/ruckus/ruckusing-migrations">Ruckusing</a> or <a href="http://www.phinx.org">Phinx</a>. If you are using a framework, try leveraging their Migration feature if available. It makes it rediculously easy to do these sort of things.

I highly advise against storing images in a database.  Store the file name reference or you will have huge bloat and decreased performance.  Use the file system, it's much better at handling that sort of thing.  For data, use the http://dev.mysql.com/doc/refman/5.5/en/optimizing-innodb-bulk-data-loading.html rather than PHP.  Check out http://www.mysql.com/products/workbench/.

A even better way than csv is to create bulk insert queries and multiple update queries and dump them in a sql file. Then source the file in mysql. This method should be even faster than creating csv.

Best way is LOAD DATA INFILE I was used it for daily upload  about 2 TB of Data into MySql this is best & optimize way to upload heavy data . I was create a php program which extract rar files into directory then Upload these files which is in form of comma separated (log files) into MySql DB . I was testing using all techniques like bulk insert into db, upload data into csv format table then modify table format from csv to myisam .