You are here:  » some xml feeds not importing

Support Forum



some xml feeds not importing

Submitted by Kiddyalley on Thu, 2011-06-30 10:48 in

i am using tradedoubler and exported 4 feeds, the xml looks similar at a glance. two feeds registers and imports fine but for some reason the other two does not autodetect and no mapping is visible in step 2. also not all products in the xml files are imported

thanks

Submitted by support on Thu, 2011-06-30 11:06

Hi,

Welcome to the forum!

Where feeds are from the same network and there seems to be an auto-detection issue with some but not all feeds, the first thing to try would be first to go Feed Registration Step 1 for one of the working feeds, and then COPY the Auto-detected Format String from that feed. Then click Register alongside one of the feeds that does not auto-detect, and PASTE the Format String from the working feed into the box, and then click Next.

If that doesn't let you register the feeds that are not working, if you could post the installation URL and any password required for /admin/ (I will remove these before posting your reply) and let me know the filename of one of the feeds that is not working I'll take a look for you right away.

Regarding not all products being imported; the criteria for a product to import are

a) Valid Name, Buy URL and Price fields
b) Unique to that merchant

In most cases where fewer products than expected are imported, it is generally down to b), for example in clothing feeds where a product name might just be "Jumper", with more information provided in another field, e.g. "COLOUR". In this particular example, Text After or Text Before filters can be used to combine the fields.

If you're not sure of course just let me know.

Hope this helps!

Cheers,
David.
--
PriceTapestry.com

Submitted by Kiddyalley on Thu, 2011-06-30 11:56

hi david
{link saved}

the tesco.xml file only shows a few products but there are suppose to be thousands.
the elc and mothercare files do not work at all, I tried pasting the auto setected string from the Halfords feed

thanks

Submitted by support on Thu, 2011-06-30 12:58

Hi,

What it looks like is a scenario that I come across very occasionally whereby the version of the underlying PHP XML parser (or more accurately the XML libraries installed on the server) can abort parsing in the case of character encoding inconsistencies.

I notice that your feeds are ISO-8859-1 encoded, and I have a work-around version of the parser that cleanses data to the specified encoding during import which I will email to you now...

Cheers,
David.
--
PriceTapestry.com

Submitted by Kiddyalley on Thu, 2011-06-30 13:10

thanks david, i will give it a try

will it help if i use a different encoding?

Submitted by support on Thu, 2011-06-30 13:14

If you have the option when generating feeds, I would suggest UTF-8 out of preference as this is the default $config_charset in the distribution version of config.php

In general, this setting should match the encoding of the majority of the feeds you are expecting to work with; and then if there are circumstances where you have no option but, say, ISO-8859-1, then UTF8 Encode (ISO-8859-1 to UTF-8) filters against the Product Name and Description fields can be used to convert the encoding at import time...

Cheers,
David.
--
PriceTapestry.com

Submitted by Kiddyalley on Thu, 2011-06-30 13:52

excellent, changing the XML to UTF-8 did the trick, thanks for the help

Submitted by brentmitchell on Mon, 2011-08-29 20:17

hey david,

i'm having a different issue, but kinda similar. i have a feed that times out and creates a 404 error when trying to import it. it doesn't show that it completed the import as it doesn't show any items listed, but the products are listed when i do a search by merchant. how can i get the this merchant to import correctly? please let me know if you have any ideas! Thanks.

OnSale-Product_Catalog.xml is the merchant in question...

Submitted by support on Mon, 2011-08-29 20:31

Hi Brent,

404 is a interesting one as normally an abort of an XML import is a timeout issue, which can be either server (e.g. Apache) or PHP enforced but these generally result in server errors rather than 404 (which is not found).

This sounds to me like a proxy server somewhere between you and your website (maybe at your ISP) is actually where the timeout is occurring so I'll email you a version of the web based import script that continuously generates output during the import which should help...

Cheers,
David.
--
PriceTapestry.com

Submitted by brentmitchell on Mon, 2011-08-29 20:55

works perfectly...thanks!