i have read most of the related postings and answers allready ... (e.g. http://anitagraser.com/2011/03/07/how-to-specify-data-types-of-csv-columns-for-use-in-qgis/ and Changing CSV layer attribute value from text to number in QGIS? ... among others)
but (!) unfortunately none of them solved my problem. as someone postet on 2012-01-04 12:27 definition of properly defined *.csvt doesn't effect field-import - they are expected to be numbers and still are string.
her is a sample of my csv
InputID,Mean,Stdev,min,max
1,334.7617168,36.87450474,294.4334857,383.5583136
2,446.3946568,207.949905,153.2494524,613.3182137
and the csvt
"Integer","Real","Real","Real","Real"
.. i have already:
- checked field names (no blanks), checked separators, ... which are qgis' default because the csv is distance matrix calculation result
defined the csvt with simle "Integer","Real" and the precision-type-style "Integer(6)","Real(8.3)"
changed country settings in win7, win8, osx 10.9 due to known decimal separator issues
- used csv as specified as well as "excel-style" csv
(";"-separated) - checked all that in qgis 2.0, 2.2, 2.3(master)
and: at the beginning this problem was related to distance matrix output
... where qgis calculates the results (and writes the csv-outputs), which of course are numbers (mean, stdev, min, max)
so, initially i was just looking for a straight forward way to use qgis' own calculation results for mapping of distance-distribution ...
is there a way to solve or completely bypass the csv-t issue in a way, i can easily communicate to gis-beginners ?


