Does WordPress’ WXR system have bugs in handling tags? (#)
WordPress provides two ways to export your data (for moving to a new server, for example). The easiest option would be to use the WordPress export functionality (WordPress eXtended RSS or WXR) from within the administration panel (Manage > Export). Another option would be to get your hands dirty (slightly) and export the database and later import it at your new host. The main disadvantage of WXR is that it does not save data created by plugins.
I had recently moved from Dreamhost to Doreo (find out why I like Doreo), and since the database export feature was throwing up some weird error, I decided to use WXR.
Upon importing the WXR file into my new host, I found that tags were rendered useless. Some tags were lost, while many others were duplicated (and named 1, 2, 3 etc.) creating a big mess of duplicate content. I cleared the database and re-imported, but tags were still not proper. Later, I tried it on a test blog, and the result was the same.
The following were the steps I took as part of damage control:
- Delete all tags
- Block the entire /tag/ directory using robots.txt (drop me a comment if you are interested in knowing how to do it).
Of course, this resulted in a rather large number of 404 ‘Not found’ errors, but it was important for me to not have multiple tag pages with same content. Now that I have settled at my new host (and decided that I would not have to move any time), I might re-tag posts (around 200 posts) manually, which is cumbersome, to say the least.
Have you had any such problems while importing WXR file?