Geek is not Dork

I felt sure that someone must have already written a web page about how to get the permalinks to transfer from the new Blogger to Wordpress. All I could find was information about Wordpress version 1.

For starters, why doesn't wordpress do it automatically, or as an option ? It does such a good job already in the import code, why not this little final step ? Seems strange to me, especially since I can see the bug report in their problem tracking system.

Anyway, enough whinging about that, here's how to get it to work.

I started with Maintain permalinks moving from Blogger to WordPress from justinsomnia.org.

Sadly the import code - blogger.php is quite different in Wordpress V2.2.1. But fear not, you only need to add just 2 lines of code and you are done.

Open the file

wordpress/wp-admin/import/blogger.php

and look for the line containing the string

wp_insert_post

- at around line 520. Insert these 2 lines after the call to wp_insert_post;

preg_match("/([^\/]+)\.html$/", $entry->old_permalink ,$matches);
$wpdb->query("UPDATE $wpdb->posts SET post_name = '" . $matches[1] . "' WHERE ID = '$post_id'");

Upload the file you have edited to your wp-admin/import/ directory.

I like having .html at the end of my URLs, so this method will create URLs on your new site that end in .html. If you don't want that, you can slightly modify the first line as ;

preg_match("/([^\/]+\.html)$/", $entry->old_permalink ,$matches);
$wpdb->query("UPDATE $wpdb->posts SET post_name = '" . $matches[1] . "' WHERE ID = '$post_id'");

How it works: after the new database record has been written for the post itself, we go back into the posts table and write a new value for `post_name'. This value is gleened from a string we got out the permalink name that was read from blogger in the first place. In effect we truncate everything up to the last forward slash in the permalink name from blogger and use that as the post-name (or slug as WP likes to call it) for the new database record.

eg; http://12testing12.blogspot.com/2007/07/an-on-and-other-articles-it-if-then.html became http://newblog.com/an-on-and-other-articles-it-if-then.html whereas it would have otherwise become http://newblog.com/a-an-on-and-other-articles-it-if-then-when.html. You have to look hard, but the first "a" is missing and an extra "when" at the end is missing also.

Blogger will normally drop articles and trucate the URL at 40 characters. Without the ability to exactly match the URI part of the address from the old blog to the new blog one could not use a simple mod_rewrite to redirect hundreds of pages. I gave a poor example, sorry, but it does make a huge difference if you want to keep your pagerank and hair.

I tried this on a blog of 295 posts and so far it has worked a treat. Let me know if it works for you. I haven't proved that the mod_rewrite works yet, but it looks like it should !