Google Base Feeder Support Thread [OLD]

simon1066:

**Found another reason why some products not exported to XML file.
**
For some time I have been unable to get all of my products included in the XML file. I realise that there are a large number of factors influencing this and I think I have taken all of them into account.

The only way I have found to get one of my new products included is to temporarily reduce its product name to a maximum of 28 characters and edit the exported XML back to the original name length.

The problem is it seems to be only relevant to that one product. i.e. renaming an existing product with a shorter name to >28 char still works. Can't figure it out. One thing that seemed to make a difference initially was how the new product was generated - creating a new one or copying an exisitng one. Can't seem to replicate this though. Additionally the missing product didn't appear in the debug file at all.

Anyway, I would be interested to hear if anyone else has resolved the XML inclusion problem in the same way. In the meantime I'll try to replicate the problem more thoroughly on my local copy.

Cheers for that Simon! I started encountering this problem when i upgraded from 1.8.2 to 1.10.3 and shortening the title to less than 28 char fixes it!

Here's the killer though...

The two products in question have the same title but different MPN's, id no 8 gets included in the XML regardless while id no.9 only gets included if i shorten it's title to less than 28 char :blink:

I suspect that as the two products are very similar, id9 may have been copied from id8.

When i upgraded from 1.8.2 to 1.10.3 i converted the db to utf-8 following the instructions provided, i wonder if it has anything to do with that. The product descriptions also struggle with copyright and trademark symbols now too, i have to switch to source <> and replace them with the character entities to avoid encoding errors when i upload the feed.

RescoCCC:

You may be confusing the attribute Product Type with Google Product Category. Leave Default Product Type empty, set Product Type to "top", and fill in Default Product Category with the correct entry from the Google Taxonomy.

Except there is no "default product category". I applied the update, even uninstalled the script and reinstalled it from scratch. The default product category shows to be in the database but not in the front end configuration.

Well... this is a first for me! Did a test submission of one of my clients' xml feeds into GoogleBase...

... and got NO WARNINGS! :clap:

hardwiredtemplates:

Delia I think we found our answer. I went to version 1.10.2 and the feed does load.

Still wondering if Adding these mods would have helped.
Hi, hardwiredtemplates ... whether or not conversion is really necessary, the Latin 1 to UTF8 does work, I've run it against a localhost test file, and have not found any problems with it so far. It changed the collation for the database, each of the tables, and all the pertinent fields in the tables to utf8_general-ci. Dr Byte has another feather in his cap.

I think the tutorial instructions say to change the collation to utf8_unicode_ci though, and I don't know if that's a killer or not. Also, have no idea if it's really necessary to change the collation, though I get the feeling most people don't. If someone knows, would appreciate the input.

Hello, all! We got into working on this a couple of weeks ago (I know that's late, but ... well ... things happen). So far we've installed and tested against a localhost test company, and mostly the GMCF seems to be working, with one big issue, at least for us.

Trying to run against all products, somewhere along the line the feed chokes and displays

Google Merchant Center Feeder v1.10.2 started 2011/09/24 14:48:11
Feed file - C:/wamp/www/ct139googlebase/feed/google/CT_base_google.xml_products_en.xml
Processing: Feed - Yes, Upload - No

with no record counts. Looking at the XML, it ends with an item, has an /item> tag, but no </channel> or </rss> tags. Cutting down on the number of records to around 600-800 by limiting the manufacturers included makes it work, and Google seems to like it, the test feed shows Google inserts all the records (with a Google product category warning, of course).

I've read of similar symptoms here, but can't find whether anyone has a solution, or even an indication of what the problem might be. If you can shed some light on this, it would be much appreciated, thanks in advance for any input.

Installed the new Google Merchant Centre module and after some initial issues (a bit of advice to anyone with Google Base Feeder module - uninstall that before adding this module) it has worked fine. Except that is for one issue.

The feed runs fine but when clicking on the link we encounter the following error 'error on line 14 at column 54: EntityRef: expecting ';''. Now after a fair bit of investigating we found that the brand name S&C was causing the problems. Specifically the s&c characters in the image url was resulting in the xml feed being rejected. For the time being we are having to download the feed amend the '&' using '%26' before manually uploading the feed. As you can imagine that is not viable in the long run.

My question is that we did not encounter this error when running Google Base Feeder so what has changed to cause this issue? And is there a fix for it? Any help would be appreciated.

You can view the generated feed here

http://www.123av.co.uk/feed/google/123_froogle_products_en.xml

And we're running version 1.3.9h of Zen Cart.

jxl1:

Installed the new Google Merchant Centre module and after some initial issues (a bit of advice to anyone with Google Base Feeder module - uninstall that before adding this module) it has worked fine. Except that is for one issue.

The feed runs fine but when clicking on the link we encounter the following error 'error on line 14 at column 54: EntityRef: expecting ';''. Now after a fair bit of investigating we found that the brand name S&C was causing the problems. Specifically the s&c characters in the image url was resulting in the xml feed being rejected. For the time being we are having to download the feed amend the '&' using '%26' before manually uploading the feed. As you can imagine that is not viable in the long run.

My question is that we did not encounter this error when running Google Base Feeder so what has changed to cause this issue? And is there a fix for it? Any help would be appreciated.

You can view the generated feed here

http://www.123av.co.uk/feed/google/123_froogle_products_en.xml

And we're running version 1.3.9h of Zen Cart.

Hello, jxl1 ... you don't mention which version of GMCF you installed. I understand that in 1.10.3 the sanitizer code has been removed, which might resolve that. 1.10.2 does have code in it, that you can probably amend to fit your needs. That is in /includes/classes/google_base.php I think.

inovermyhead:

Hello, jxl1 ... you don't mention which version of GMCF you installed. I understand that in 1.10.3 the sanitizer code has been removed, which might resolve that. 1.10.2 does have code in it, that you can probably amend to fit your needs. That is in /includes/classes/google_base.php I think.
The problem is that google does not allow ampersands in the image name and the santizer does not change that so therefore the problem. Just remove the ampersands from the images - you should never use symbols in your image names. Make the image names with keywords- words that matter.

You're right, Delia, didn't read it right, and didn't realize it was an image. Sorry.

delia:

The problem is that google does not allow ampersands in the image name and the santizer does not change that so therefore the problem. Just remove the ampersands from the images - you should never use symbols in your image names. Make the image names with keywords- words that matter.

inovermyhead - you mentioned a few posts back, that you converted a database from latin1 to utf8 and didn't see any problems.

Using drbytes converter I converted a 1.3.9d database from latin1 to utf8 and after a short while I realized many (not all) descriptions were truncated. Did you notice anything like that happening?

buildingblocks:

inovermyhead - you mentioned a few posts back, that you converted a database from latin1 to utf8 and didn't see any problems.

Using drbytes converter I converted a 1.3.9d database from latin1 to utf8 and after a short while I realized many (not all) descriptions were truncated. Did you notice anything like that happening?

Hello, building blocks. Not at the time, just did a random check of several listings.

Went back a bit ago and checked further, especially on listings that have a special character in them, like apostrophe, double quote, or an accented character, and as you say, the listing is truncated immediately after the special character, and in one case the sideboxes appeared after the description, rather than alongside, indicating the HTML can be affected as well. :shocking:

Using 1.3.9h

Thanks for clarifying that inovermyhead. At least I know it's not just me having done something wrong.

Thanks for all the responses it is much appreciated.

delia:

The problem is that google does not allow ampersands in the image name and the santizer does not change that so therefore the problem. Just remove the ampersands from the images - you should never use symbols in your image names. Make the image names with keywords- words that matter.

With regard to this the exact same product was being accepted when we were running Google Base Feeder, so it can't simply be a case of Google not accepting the format.

Also while it would be great if we could remove the ampersands it is automatically generated by Ceon SEO URL module we have installed and comes from the brand name S&C.

With regard to this the exact same product was being accepted when we were running Google Base Feeder, so it can't simply be a case of Google not accepting the format.

It's likely that the version you were using had the sanitizer.

Google finally did change the messaging and has posted a new blog post that says:
"Starting September 22, we'll gradually begin enforcing the new requirements by removing items from Google Product Search that do not meet the new requirements and, in some cases, suspending accounts that do not comply."

So the 22 deadline wasn't as hard as we anticipated it would be.

for some the deadline came down hard.

I saw this one yesterday.

"All your items have been suspended from the destinations and countries listed below. Please see the Data quality section for details."

buildingblocks:

for some the deadline came down hard.

I saw this one yesterday.

"All your items have been suspended from the destinations and countries listed below. Please see the Data quality section for details."

Okay, thanks for the report!

They have improved the messages about problems for some - not all customers seeing the new look. Just ran into a section titled: Critical Errors: Must be fixed to prevent account suspension in United States.

That's much clearer than before.

Feeds are processing fine on all sites with the new version. No news on the clothing attributes that are now required. The feed and items are still listed. However we have not "resubmitted" the feed. Probably get kicked out when they review it? Just missing color, size, etc... No critical error messages yet. We'll see what happens?

Okay, my first disapproval -
the cart has linked products which means duplicate products to Google. I hadn't seen this before but had just changed to show all products regardless of stock quantity.

I had wondered about this before with another customer and I'm not seeing how to exclude those addtional linked products.

I had also changed the product type in admin to top where I had been using full - path.

Don't know what to do here - anybody with experience with this scenario?