[Qgis-user] Delimited text plugin with fields containing commas

[Alex Mandel]

If it was being entered on the web it should have gone straight into a
database. So there should be no need for a text file, it should all be
db dumps if you need to move things. Personally I'd recommend putting it
in SQlite or Postgres to start then you can move directly to Spatialite,
Postgis - or use those from the beginning.

It just raises the need to really think about data entry in project
design and to get good at writing fancy python scripts to clean up bad
text files people give you based on regex rules and other magic.

Enjoy,
Alex

On 06/21/2011 12:48 AM, M.E.Dodd wrote:
> I've had a whole range of similar problems with a large database of publically entered records from many countries, it contains all sorts of strange characters (in many languages) including most characters you'd think of as possibilities for column separators.  Rather a nightmare to tidy up and analyse, end up stripping out loads of different characters just to get it to read by spreadsheet and gis.  Ideally there might be some completely different separator that could easily be edited in to show the columns and keep most of the commas and other characters within the columns but the big issue is how to do that editing in the first place to correctly identify the columns or easily allow moving of text between columns if an automatic import gets it wrong.
> I have sorted most of it out now, in my case, by long laborious means but if someone could come up with a good way of dealing with this kind of messy file (entered by general public in many countries so with potentially unpredictable strange characters) that would be very useful.  Just before you say you should have been much more restrictive on the web input, we were fairly restrictive but still need to allow quite a range of possible inputs in free text in any language.
> 
> From: John Callahan [mailto:john.callahan at udel.edu]
> Sent: 21 June 2011 00:36
> To: tech at wildintellect.com
> Cc: qgis-user
> Subject: Re: [Qgis-user] Delimited text plugin with fields containing commas
> 
> You're correct.  That way probably would be the preferred work-around.
> 
> - John
> 
> 
> On Mon, Jun 20, 2011 at 5:18 PM, Alex Mandel <tech_dev at wildintellect.com<mailto:tech_dev at wildintellect.com>> wrote:
> I agree quotes should work but I've found many parsers to not follow the
> expectation on this. As for semicolons I only meant as the delimiter
> leaving the commas inside your text. That way you can tell the parser
> that ; is the separator between records.
> 
> Thanks,
> Alex
> 
> On 06/20/2011 01:04 PM, John Callahan wrote:
>> I use semi-colons when I can but have run into situations where commas are
>> necessary, such as names of places.  I agree with the work-around and I've
>> done that before.   As long as quotes (") are included around the values, it
>> should work, and I believe it was working for a while.
>>
>> - John
>>
>> ***********************************
>> John Callahan, Research Scientist
>> Delaware Geological Survey
>> University of Delaware
>> URL: http://www.dgs.udel.edu
>> *******************************
>>
>>
>> On Mon, Jun 20, 2011 at 3:59 PM, Alex Mandel <tech_dev at wildintellect.com<mailto:tech_dev at wildintellect.com>>wrote:
>>
>>> On 06/20/2011 12:35 PM, John Callahan wrote:
>>>> Has anyone seen this problem with the Delimited Text plugin?  I am seeing
>>>> this in today's download of QGIS 1.7 standalone on Windows, and on a
>>> recent
>>>> install through OSGeo4W of 1.8-trunk.
>>>>
>>>> "Delimited text" plugin doesn't allow to load csv file with field with
>>>> commas
>>>> http://hub.qgis.org/issues/2208
>>>>
>>>> - John
>>>
>>> I have had that problem before, with lots csv import tools (not just
>>> qgis). Are you using commas to separate the values too? I usually have
>>> much better success changing that to ; or |
>>> so instead of "2","test,test","1"
>>> "2";"test,test";"1"
>>>
>>> Easiest way to swap out the delimiter is to use OpenOffice/LibreOffice
>>> and change it when saving.
>>>
>>> Thanks,
>>> Alex