Since there’s so little information on the internet about exporting Blogsome blogs to WordPress, I thought I’d write up a quick guide based on my experience. (Regular readers, feel free to hit the eject button here.) This guide is specifically for people who are working with Japanese (or other funky languages). It might be useful for people who are trying to move their blogs but are having trouble getting their characters to look right. When I first imported my SQL, all the Japanese text had been replaced with question marks. Yikes. The problem it seems was that while the input on Blogsome (and WordPress) is UTF8, the databases themselves are automatically set to Latin1 for some reason. (In my case, Swedish Latin1, I might add!) Mixing these character sets will garble all your kanjis. Here’s how you move your blog and fix garbled chars:
1. Get an SQL dump from a Blogsome admin. This is easy enough to do. Just do some searches on the forums and email the admin once you’ve found the admin’s email.
2. Install WordPress. Get WordPress running on your site, and make sure it works properly. You can do all the formatting now or after you’ve imported all your posts. Either way should be fine.
3. Use phpMyAdmin to export your SQL for the just installed WordPress.
4. Delete the SQL database for the WordPress install. I did this in the cPanel provided by my host.
5. Create a new database with the same name as the one you just deleted (in cPanel), and change its character set to UTF8 under phpMyAdmin. It will have no tables in it.
6. Open the SQL files you just exported in a text editor and find and replace all cases of “latin1” with “utf8”. Make sure you also save the file itself in UTF8 format. I used TextWrangler, which is an OSX program with the great option “Reopen in UTF8” or any other character set you might need.
7. Import the SQL back into your database. The blog should be back to where it started (default status with one test post), but now the SQL database and all your tables are set to UTF8.
8. Split your Blogsome SQL file up into individual files for each table. This blog post was an enormous help to me. Basically, the SQL breaks down into individual tables for different sets of information like comments, options, post data, etc. Blogsome is based on WordPress, but WordPress has been upgraded several times since Blogsome started, and not all of the tables are the same, so this is an essential step. Use the table name as the file name. You’ll end up with posts.sql, comments.sql, postmeta.sql, and a bunch of others.
9. Find and replace all cases of “latin1” with “utf8” in your new SQL files. This will ensure that the tables will be properly imported in the UTF8 character set. Again, don’t forget to save the files themselves in UTF8 as well.
10. Import the necessary Blogsome SQL files. You won’t be able to import all of them, specifically the categories and linkcategories tables, since those have been combined into the terms table. I only imported 5 SQL tables total: posts, comments, users, postmeta, and links.
11. Open the tables you imported in phpMyAdmin, and edit the table values so that they match the database descriptions listed on the WordPress site. I’m not quite sure exactly what they all mean, but I’m assuming they’re just instructions about what kind of data you’ll be inputting into the database. You’ll have to add and delete some rows of the table.
12. Open your blog. Everything should be working. For some reason a few links would not register on my blogroll but were still registered in the WordPress admin page. All I had to do was open them in edit and click update, and they came back. Also, the number of comments won’t appear on my front page, but the comments themselves are all still there. If one comment is added on top of the existing comments, then the correct number appears. Other than that, I haven’t had any problems (*knock on wood*).
Hope somebody finds this useful!