We have two students whose names differ only by a single tone. Their given names are Cú and Cu, and the rest of their names are identical. When Data Admin imports Markbook Data using “Surname, Preferred Name” to connect the data to the proper student, it ignores diacritics. Sometimes that is helpful, but the result is that Cú’s grade goes to Cu, and Cú does not receive a grade. We have to manually correct them.
When I first started using Data Admin, I was surprised that, in contrast to other kinds of data imports that use the student’s username, in this case Data Admin asks for “Surname, Preferred Name”. Then I reflected on who would provide such data (teachers) and decided that made sense. But in certain cases it does not work, like in the case above.
Can I suggest adding the option to import based on username?
Sure, this sounds quite reasonable and is a good example of extending the system to work for a variety of needs.
You’re right, when I setup that import type I wanted it to be end-user friendly (as much as possible, being spreadsheet data) so I went with student name over username. Normally this wouldn’t be precise enough, but in the back end it’s cross-checking both name and class enrollment to enter marks for the appropriate student. In your case are the students with similar names also in the same classes?
I suspect the comparison between Cú and Cu resulting in a match is related to using MySQL utf8_general_ci collation, which seems to ignore not only case but also accents. This could certainly have some interesting implications for overall internationalization, so its something to keep in mind for development down the road.
Here’s an import definition you can use, I’ll add it to the next update for the module. For now you can upload it to either /uploads/imports or <code class="CodeInline">/modules/Data Admin/imports. The file is zipped because the form doesn’t like YML files, so unzip it first and place in either folder location (the Custom Imports folder exists as of Data Adminv1.1.0)