CSV Import feature: Low-hanging improvements?

Hello everybody,

yesterday I added quite some gray hairs to my many by doing a very simple address import to a new Tryton instance. There has been lots of discussion on improvement on the CSV importer, but nothing really happened, AFAIK. For that reason, here are some suggestions which may be easy to implement, but should achieve considerable improvement:

  • Provide a better message on invalid phone numbers. 0xxxx is a valid number format in Germany (maybe in other countries), but seems no longer be accepted by Tryton. Now it seems to regard only “+[country_code]xxxxx” as valid. It took me 1 1/2hrs to find that out, a simple message “please provide phone number in ‘+[contry_code]xxx’ format” would help a lot.
  • The importer cancels after the first invalid input it finds. So when you are not really familiar with the data structure required, you will do many circles - as I did again yesterday. So it would be great if the importer could walk through the whole import file and provide a list of invalid inputs (notification or file), so the user can correct all of them at once.
  • In Germany, we cannot use ‘,’ (comma) as a CSV separator, because we use that one as a decimal separator. ‘;’ (semicolon) is perfect, but it’s really annoying having to change the CSV separator at every repeated attempt. It would be great if Tryton could remember the CSV settings once changed.

Sorry - I cannot provide a patch for these things, I don’t have coding abilities. Shame on me.

But I’d be happy to offer a bounty of 50€ for these improvements, hoping that other fellows will join in with small offerings as well, so a reasonable reward for these improvements will be gathered.

No need to mention that the import feature is one of the entrance gates to Tryton for newcomers. So in case Tryton is interested to acquire more users, here would be an effective lever.

EG

Format without country can still be used as long as there is an address with this country on the party.
We can not guess the correct format of an incorrect phone number.
Also the library used to validate the phone numbers provides only some detail but in English.

This is because data may be related so without the failing row the others may also be failing and an invalid column may make other columns invalid.
And also because it is much expected that if a row has a validation problem, all the others have the same problem. So it will not be helpful to raise an error message for hundred or thousands rows for the same error.

You can quote the content with comma.