Hi there,
I couldn’t find this listed elsewhere on the forum, but I just noticed that .csv exports of text fields that include vulgar fractions (the pretty unicode symbols of fractions like 1/2 = ½) do not export properly due to an encoding issue when exporting to .csv, even though they display properly in the front end database.
Description:
Unicode vulgar fractions (and probably other unicode symbols) do not encode properly in .csv exports. Visible using this query when signed into the ohiomollusk database.
Steps to reproduce:
Query for any of the following symbols in a text box, with “distinct” unchecked: ⅓, ⅔, ⅕, ⅖, ⅗, ⅘, ⅙, ⅚, ⅛, ⅜, ⅝, ⅞
Click “create .csv” (note: this error does not affect .kml files, the symbols are reproduced correctly in exported .kmls)
We will probably convert these values to normal text representations of the fractions, but it seems worth reporting and fixing since a lot of software automatically converts fractions to these unicode symbols, and these numbers are essential information that get mangled in exports currently.
Specify 7 should be able to either handle this on export OR block users from entering/uploading text that includes these symbols into Specify.
What is the program that you are opening the resulting csv file in? If it is excel, does the problem persist if you use Data > From Text/CSV and select UTF-8 encoding for the import?
If you open the csv in a text editor, how does it look?
Aha! You’re onto something. When opened in moderncsv, the symbols are correct. So this might just be an excel problem.
When importing using Data>get data>from text/csv the correct conversion is automatically applied and the symbols are shown.
But the default dialog when opening the file in excel normally (using default settings) the data get garbled. Assuming this isn’t a bug, and is the expected behavior, is there a way (by changing how the .csv is formatted, or by changing a setting in excel) to have excel play nice by default? Edit: I should add that the main thing I’m worried about is end users of our data. I want them to have to do as little as possible.
