[Qgis-user] csv import issue

[Nathan Woodrow]

I would recommend using the Load Delimited Text Layer plugin (in 1.8, core
feature in 2.0) because it will handle types automatically.

- Nathan


On Sun, Sep 8, 2013 at 12:30 AM, Telmo Nunes <pinanunes at gmail.com> wrote:

> Olá Hugo,
>
> Thanks! csvt can solve the problem!
> Now a question remains: Why then having the "Quote" option on the csv
> loading dialog? Just to remove them (that would save the find/replace
> solution from your first link) ?
> On the "user friendly" prespective (what Qgis aims I believe) it could be
> used to check quoted columns and create the csvt file with text for those
> columns before loading it (or since we there, make a simple dialog to pick
> the format for each field and create the csvt file before loading the csv).
>
> Creating the csvt is not an issue (for some users), but is an extra
> "manual" processing step that I believe could be avoided.
>
> Anyways thanks for the tip!
>
> Cheers
>
> Telmo
>
>
> 2013/9/7 Hugo <hfpmartins at gmail.com>
>
>> Ola Telmo,
>>
>> OGR supports an additional file (.csvt) to specify the types of the
>> attribute columns of a csv file. Check these blogs:
>>
>> http://sigsemgrilhetas.wordpress.com/2013/05/07/visualizar-dados-dos-censos-2011-com-o-qgis/
>>
>> http://anitagraser.com/2011/03/07/how-to-specify-data-types-of-csv-columns-for-use-in-qgis/
>>
>> Cheers,
>>
>>
>>
>> On Sat, Sep 7, 2013 at 1:20 PM, Telmo Nunes <pinanunes at gmail.com> wrote:
>>
>>> I have a recurrent issue when loading csv files.
>>> When one of the fields is a text field only with numeric characters qgis
>>> allways imports that field as numeric. You might ask why would I want a
>>> field with numeric characters to be a text field, and the answer is that
>>> one dataset I usually use, the administrative regions regions are coded
>>> with numeric characters as text...for example 011312 that is converted to
>>> 11312 even if the values in the csv file have the quote markers (" or ')
>>> and I define that on the csv import dialog (see below a csv example)
>>>
>>> I have found diferent workarround solutions for this issue like:
>>>
>>> - converting the code on the original shapefile to numeric
>>> - importing the csv to a database and loading from the database
>>> - creating a dummy record where i put a letter character in that field
>>>
>>> They all work, but is just annoying to do this over and over and adds
>>> some unecessary steps in the data analysis workflow (for example when
>>> importing a satscan gis file output to map the clusters). My question is:
>>>
>>> - Am I missing something here?
>>> - If not, is it possible to fix that?
>>>
>>> Below is a portion of a csv file, if you manage to import the file and
>>> keep the LOC_ID field as text please let me know.
>>>
>>>
>>> 'ID';'LOC_ID';'CLUSTER';'P_VALUE';'CLU_OBS';'CLU_EXP';'CLU_ODE';'CLU_RISK';'LOC_OBS';'LOC_EXP';'LOC_ODE';'LOC_RISK'
>>> 1;'070105';1,00;1,0e-17;283,00;72,06;3,92;13,44;0,00;0,00;0,00;0,00
>>> 2;'070101';1,00;1,0e-17;283,00;72,06;3,92;13,44;0,00;2,58;0,00;0,00
>>> 3;'070103';1,00;1,0e-17;283,00;72,06;3,92;13,44;0,00;2,72;0,00;0,00
>>> 4;'071002';1,00;1,0e-17;283,00;72,06;3,92;13,44;47,00;3,12;15,03;17,07
>>>
>>> My best regards
>>>
>>> Telmo Nunes
>>>
>>> _______________________________________________
>>> Qgis-user mailing list
>>> Qgis-user at lists.osgeo.org
>>> http://lists.osgeo.org/mailman/listinfo/qgis-user
>>>
>>
>>
>>
>> --
>> Hugo Martins
>>
>
>
> _______________________________________________
> Qgis-user mailing list
> Qgis-user at lists.osgeo.org
> http://lists.osgeo.org/mailman/listinfo/qgis-user
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/qgis-user/attachments/20130908/67954624/attachment.html>