Home » Forums » Ubercart » Announcements

What to do with the Importer

Submitted by Lyle on Thu, 05/29/2008 - 10:13
Lyle

Posts: 2267
 Joined: 08/07/2007
AdministratoreLiTe!

The more people use the Importer module, the less useful it seems. For the most part, I've been concentrating on other parts of Ubercart rather than trying to improve or fix the importer, but now that the 1.0 release is imminent, I'm not as sure what to do about it.

My inclination is not to include it or the Repeater module in the 1.0 release, but add them as contrib modules on d.o. This lets those of you who are actually using it to keep on using it, but it's one less thing for those who don't to worry about. And it lets us release Ubercart 1.0 that much sooner.

What say ye all?
permalink #1 Thu, 05/29/2008 - 10:19
codexmas@drupal.org

Posts: 34
 Joined: 09/19/2007

Sounds good. I have found writing my own custom importers much easier to manage a usually one time operation. I have also written an updater, quite client specific, that could be part of a discussion on how to make some great contrib modules. Eventually after enough love is given to these needs they may be core worthy.

Cheers!
Gord.

permalink #2 Thu, 05/29/2008 - 10:28
Ryan

Posts: 5367
 Joined: 08/07/2007
AdministratorHead Code Monkey - I eat bugs.

I think I'm in favor of this... I just wish it wasn't so late in the development cycle. I'm torn between being able to say that we do in fact have an importer, even if it's limited in its functionality, and having to tell people it's available through a contributed module. Some folks get freaked out by contributed modules for some reason.

Anyways... if this module is full of irreconcilable bugs, I'd rather it not hold up the entire project from moving forward. If it's just not feature complete, that's another story. I think it's alright to ship something that isn't feature complete but is still usable for what it is. (Example: roles or stock module; these aren't do-all modules, but they work for what they do.)

I do think if we're going to be serious about pulling folks in from other systems we need to have a good way to get their data into Ubercart. Maybe Lyle's learned a few things since the original module that would facilitate this in a rebuild. Also, we're looking to go PHP 5 once we migrate to Drupal 6, so maybe that will help some? Or was the importer already dependent on PHP 5?

permalink #3 Thu, 05/29/2008 - 11:31
futurist

Posts: 54
 Joined: 08/08/2007

uc_importer was very important to me to migrate hundreds of products from a different shop solution to Ubercart. (I had started to write my own importer back in Ubercart's early Alpha days, but had to give up because it meant investing so much time into learning both Drupal's and Ubercart's API and DB structure. I preferred to concentrate on getting the data from an Access DB to uc_importer compatible XML. It turned out that after an update to Ubercart Beta 7 I had to rewrite some of my migration code and importing wasn't completely smooth but the module did do the job!)

Ubercart needs the importer. If debugging the current module is holding up release 1, I'm in favor of making it available as contributed module (contributed by Lyle...?) then rebuild/debug it and include it with Ubercart in a future release. Otherwise, if the module just isn't feature complete, why don't you go ahead and include it in release 1 along with an up to date documentation page that informs what the module can and cannot do.

permalink #4 Thu, 05/29/2008 - 11:33
goodeit
Posts: 107
 Joined: 05/28/2008

To me, the importer is a pretty integral part of the program. Without it, it is impossible to migrate from an existing site, or even to use any old data at all. Then being able to export the data out and bring it back in makes it easier to bulk change prices, keep things current elsewhere, etc. I wouldn't be against moving the importer to a contrib module as long as it doesn't get abandoned and forgotten.

permalink #5 Thu, 05/29/2008 - 11:50
mattie@drupal.org
Posts: 6
 Joined: 05/20/2008

Hi, I am a new user of ubercart and would like to implement a new webshop with it. so migrating from another existing site is not an issue for me. However, if I understood correctly, the combination importer/repeater allows a shop admin to propagate product changes to multiple sites automatically. I was amazed that ubercart supported that, as I had received such a request before but said it was most probably not possible Smiling

However, since it is a beginning shop, I don't need multiple affiliates from the beginning, so if this really hampers the v1 release, please skip it. But it would be nice if it would come back in a future release.

so, that's just my case of course Smiling

permalink #6 Thu, 05/29/2008 - 12:51
Blessed_Steve

Posts: 163
 Joined: 08/07/2007
Uber DonorBug FinderEarly adopter... addicted to alphas.Getting busy with the Ubercode.Not Kulvik

With thousands of products, we had a inventory challenge on our hands no matter what. Exporting our products from our ERP into a common XML format and feeding them into the importer gaves us a valuable head start. There is no way, imho, to avoid the custom export from whereever ones products exist into a format ubercart will accept. With out the importer it would have been an even more difficult project however.

2 cents.
Steve

permalink #7 Thu, 05/29/2008 - 13:18
Lyle

Posts: 2267
 Joined: 08/07/2007
AdministratoreLiTe!

Just to make things clear, I'm not thinking about letting the importer disappear altogether. Our parent company needs it for their restaurant equipment site, so it'll definitely be developed. I may need periodic reminders though. Eye-wink

The consensus seems to be that it can be a contrib so that Ubercart core can be given the go ahead. I'll make a project on drupal.org, and we can go from there.

permalink #8 Thu, 05/29/2008 - 13:55
Ryan

Posts: 5367
 Joined: 08/07/2007
AdministratorHead Code Monkey - I eat bugs.

Maybe hold off and let the idea sit for a day? The importer has been around this far, no need to get rid of it so fast. Also, there might be others with stronger opinions/arguments to leave it in that didn't access u.o in the 2 hours the question was open. Sticking out tongue

I'd still like to consider whether or not the code there is buggy vs. not feature complete. I don't mind advertising and documenting limited importer capabilities. From an outsider's perspective, having a freely available contributed importer vs. a built-in importer actually does make a difference... it's harder to advertise community add-on features than it is a core feature set.

permalink #9 Thu, 05/29/2008 - 15:21
dstrickland

Posts: 28
 Joined: 01/14/2008

Like others, it is a integral part of the application, but I don't care if it is core or a contributed module. You could argue that if it is core, it will get more attention and more likely work.

I have only needed it twice. Once was a corrupted database that would cause the application to crash when I would try to view or edit a particular user. The importer was integral to brining up the other replacement site. I only wish that it would carry the shippable flag on export and import, but that was easy to fix with SQL. It was just something to know for the readme.. I guess.

--David

--

David Strickland
Delante Solutions

permalink #10 Thu, 05/29/2008 - 15:42
Andy

Posts: 151
 Joined: 08/07/2007
AdministratorNot Kulvik

I agree with Ryan. If it is reasonably bug free, keep it in 1.0 even if it is not feature complete. As someone who has experience exporting and importing to and from several e-commerce solutions, I do not expect an importer to work perfectly for me. It never seems to work that way. I expect the importer to do most of the heavy lifting while I fine tune it for my specific needs. I believe the UC importer does this and therefore should be included in the 1.0 release. Just my opinion.
Peace,
Andy

permalink #11 Sat, 05/31/2008 - 11:15
goodeit
Posts: 107
 Joined: 05/28/2008

Lyle/Ryan:
For those of us who _really_ need a working import/export function in order to use Ubercart, what can we do to help? Would creating a bounty help, or possibly trying to work on it ourselves (PHP is not my thing, but I'll give it a try...)? Are bug reports helpful or hurtful at this stage?

Thanks