Feature request: Accent and punctuation insensitive search
On a song database I inherited, there are a bunch of songs that have the word blessèd in the title (note the accent on the second e), which do not appear if you search for blessed. Of course we could just edit all the titles, but that's not a solution everyone would be happy with.
There are also songs with punctuation marks in the title, most commonly commas though the odd exclamation point and other punctuation makes an appearance. Again, the search doesn't find the song unless it's typed in exactly as it appears, and I don't really have space in my head to remember which score of the two score songs starting with "Come " actually start "Come, ".
It would be nice if there were an option, ideally a default, in which the search filtered out accents and punctuation, searching purely on A-Z, 0-9 and space. I appreciate that I'm coming at this purely from an English perspective, and the ideal may be different for other languages, but I can't say what that ideal would be.
There are also songs with punctuation marks in the title, most commonly commas though the odd exclamation point and other punctuation makes an appearance. Again, the search doesn't find the song unless it's typed in exactly as it appears, and I don't really have space in my head to remember which score of the two score songs starting with "Come " actually start "Come, ".
It would be nice if there were an option, ideally a default, in which the search filtered out accents and punctuation, searching purely on A-Z, 0-9 and space. I appreciate that I'm coming at this purely from an English perspective, and the ideal may be different for other languages, but I can't say what that ideal would be.
Comments
Tomas
In practice the easiest way to do it would almost certainly be to store a stripped down copy of any data that needs to be searched this way, alongside a copy for display with everything done properly. Each search would have to start by stripping out special characters from the search string, but that's a fairly easy thing to do given the typical length we're thinking about. It may even be possible, and perhaps a little more efficient, to take care of case at this stage too.
None of that is technically very difficult. It would need care to make sure the appropriate changes are always made, and of course any changes to database formats need to be carefully thought through.
It is concievably possible that part of the task might be outsourced to the database directly without changing anything else; for example Microsoft's SQL server has a facility to do accent insensitive comparisons just as it can do case insensitive comparisons. That might not be available for OpenLP's choice of database though, and punctuation insensitivity seems to be a rarer facility still.