Authors misattributed on big MT export to WordPress

Can you create mt-export.txt in smaller chunks?

Thanks for responding Aplidji. I’ve tried that but it has no effect. Each time I have to delete entries and clean the database and start again, the whole process of even trying things out is time-consuming.

Is there a better plug in or some other way of doing it? I can see this problem goes back years. Someone must have fixed it or found a way by now surely?

Have you tried something like this?

Again apljdi thanks for the suggestion. I have just been trying to get the file splitter route to work, trying uploading via the ftp and the simple file upload options but again the import seems to be shifting author attribution, even when I have deleted every post and author except admin.

It seems like such a stupid bug and one that should have been dealt with years ago. I understand that blogs with lots of authors may not be that numerous but they might be among the most interesting.

Maybe I should be looking for a plugin that lets me reattribute authorship en masse instead?

I’m guessing that you might have two different problems. One is with file size since you are getting a a ‘crash’ partway through. That is where splitting the file should help. The second is with a possible bug in the script. I looked around some and did see some suggestions that the MT importer might not be up to date with the latest MT, but I couldn’t find enough to really be confident in that conclusion.

I think you are right. File size is causing the crashes but that wouldn’t be a problem if I could just fire it up again as it is supposed to work. I’m actually exporting from MT 3.31 (the update to MT4 or 5 being quite a big job so I was nudged into thinking about WordPress more seriously).

I have added the php line php_value memory_limit 512M

to my .htaccess file and I am getting through more entries before crashing but still the author issue continues.

I looked to see if there was a plugin that allowed me to make mass changes to authors to help me clean up the mess afterwards but couldn’t find anything.

but that wouldn’t be a problem if I could just fire it up again as it is supposed to work.

…depends on how smart the script is. Does is claim to be able to pick up where it left off like that?

There are several factors involved in files size limits, one of which is actual system level limits that PHP has no control over.

If everything else were to work, it might be your best bet to run the script and then run another to reset your authors but that doesn’t solve the file size issue. Can you post a part of that mt-export.txt to the pastebin?

it says on the import via MT page:

The importer is smart enough not to import duplicates, so you can run this multiple times without worry if—for whatever reason—it doesn’t finish. If you get an out of memory error try splitting up the import file into pieces.

Which is true apart from the author issue.

Here is some of the file https://pastebin.com/TVq0nLup

Just to make it a little trickier, the first upload has the right authorship and the subsequent ones get shuffled so it is not consistently wrong!

So you think it would work if the upload would finish in one go?

Also, have you seen this thread? … meant to post it earlier.

It ought to be pretty simple to read that file, pull out author and post name information and reassign authors, by the way.

I think it would work in one go but the file is 18MB so I am doubtful it’ll ever happen!

Thanks for the thread. I really should check my AMs, PMs and UTFs but I think that is a luxury problem to have right now!

Are you saying that a script would compare the mt-export file’s author and post name information and then apply it to correct whatever is in the WordPress database? If so how?

Yes. The author and post title are clearly labeled so getting them should be easy. I’d probably run the file through grep or sed or awk and output a very much smaller file with only those two values. Then make a kind of dummy plugin to loop through the file. You’d want to query the WP db for both author ID and post ID. If you have both, change the author, otherwise print a message saying “this post/author couldn’t be found… etc”

This is way out of my skill range. What kind of programmer would I need to find to do this? Someone with php and db experience I guess? Would they need to know about WordPress?

Anybody familiar with PHP could probably do it, but familiarity with WordPress would make it easier since you’d want to use WordPress built in functions for at least the post/author modification. Its actually not a hard problem if you can write PHP even moderately well.

Wow, apljdi, I think I have found a way that works. It is clumsy and I can’t say why it works but here goes.

I imported the first batch via the uploader (rather than the ftp route) and then I trimmed off every entry that was already in the system, saved the file again and uploaded what was left. I repeated the process half a dozen times until the upload was complete and it has kept the correct authors.

I don’t know if this is relevant but I did save as UTF-8 first (although this did not make any difference before).

It appears as if something in the process where it was going through existing entries and skipping them threw the upload off in some way.

I can’t thank you enough for spending so much time helping me out and I think I would have given up otherwise.